Showing 1–50 of 226 results for author: Krishna, R

Vision-Language Grounding as Bidirectional Concept Correspondence

Authors: Jieyu Zhang, Ziqi Gao, Luke Zettlemoyer, Ranjay Krishna

Abstract: Vision-language grounding connects language to visual content, yet most existing formulations reduce grounding to a unidirectional localization problem: given a prespecified text phrase or category name, identify the corresponding image region. This setup assumes that the relevant linguistic unit is already known, overlooking a more basic challenge in grounded communication: determining which part… ▽ More Vision-language grounding connects language to visual content, yet most existing formulations reduce grounding to a unidirectional localization problem: given a prespecified text phrase or category name, identify the corresponding image region. This setup assumes that the relevant linguistic unit is already known, overlooking a more basic challenge in grounded communication: determining which parts of the text are visually referential and how they correspond to entities in the image. We formulate grounding as $\textit{bidirectional concept correspondence}$ over an image-text pair. Given an image and its paired text, the goal is to recover all correspondences between visually referential text spans and instance-level image segments, without assuming that the relevant text spans are provided. This formulation unifies common grounding tasks, including phrase grounding, referring expression grounding, and open-vocabulary detection, by treating text segmentation, image segmentation, and cross-modal alignment as a single correspondence prediction problem. To address this task, we introduce $\textbf{ConCor-1}$, a grounding model built on top of a pretrained vision-language model. It uses learnable $\textit{bridge tokens}$ to represent candidate image-text correspondences and predicts, for each token, a text mask, an image mask, and a correspondence presence score. To train and evaluate this task, we convert diverse grounding and segmentation datasets into a unified correspondence format. Experiments show that $\textbf{ConCor-1}$ consistently outperforms baselines, improving correspondence F1 by 48% on the long-caption dataset and by 29% on zero-shot LVIS, where the large category list serves as the text input. △ Less

Submitted 7 August, 2026; originally announced August 2026.

HumanCLAW: Can Vision-Language Models Act Through a Body?

Authors: Li Siyao, Jiawei Gu, Shuai Liu, Kairui Hu, Zekun Li, Linjie Li, Chengcheng Tang, Po-Chen Wu, Ivan Shugurov, Lingni Ma, Michael Zollhoefer, Sizhe An, Abhay Mittal, Amy Zhao, Ranjay Krishna, Manling Li, Ziwei Liu, Chuan Guo

Abstract: Evaluating whether a vision-language model (VLM) can act through a physical body is challenging. The outcome of an action couples the VLM's decision with motor control. When a task fails, it is hard to tell whether the VLM made a bad choice or the motor controller simply failed to execute it, e.g., losing balance and falling. In this work, we introduce HumanCLAW, an evaluation framework that decou… ▽ More Evaluating whether a vision-language model (VLM) can act through a physical body is challenging. The outcome of an action couples the VLM's decision with motor control. When a task fails, it is hard to tell whether the VLM made a bad choice or the motor controller simply failed to execute it, e.g., losing balance and falling. In this work, we introduce HumanCLAW, an evaluation framework that decouples action decision-making from low-level execution. At every step, a harnessed, off-the-shelf VLM issues an atomic skill command, and the command is translated into a sub-second chunk of continuous full-body motion with real physical consequences, including gravity and collisions. The body can therefore act freely in the physical world, while execution-side disturbances, balance and motor errors, are factored out. What remains measurable is the model's action intelligence: its moment-to-moment choice of what the body should execute next. Based on this framework, we build HumanCLAW-Bench: 1,218 long-horizon, egocentric find-navigate-interact episodes across 41 indoor scenes. We test nine state-of-the-art VLMs and find that none solves the benchmark; the best model reaches only a 16.8% success rate. Recognizing the target is not the bottleneck. What current VLMs lack is embodied self-awareness: they lose track of their own body, failing to tell where it is, whether it has reached the goal, or whether it has hit an obstacle. △ Less

Submitted 3 August, 2026; v1 submitted 29 July, 2026; originally announced July 2026.

Comments: Project page: https://human-claw.github.io/

MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction

Authors: Jianing Zhang, Chenhao Zheng, Yajun Yang, Max Argus, Rustin Soraki, Winson Han, Taira Anderson, Chun-Liang Li, Shuo Liu, Jiafei Duan, Zhongzheng Ren, Jieyu Zhang, Ranjay Krishna

Abstract: Motion forecasting is central to visual intelligence: agents must anticipate how objects will move in order to plan actions, reason about physical interactions, and synthesize realistic futures. We argue that 3D points in world coordinates provide a general representation that is class-agnostic, view-stable, compact, and directly useful for downstream tasks. We formalize the task of goal-condition… ▽ More Motion forecasting is central to visual intelligence: agents must anticipate how objects will move in order to plan actions, reason about physical interactions, and synthesize realistic futures. We argue that 3D points in world coordinates provide a general representation that is class-agnostic, view-stable, compact, and directly useful for downstream tasks. We formalize the task of goal-conditioned 3D point motion forecasting: given a short visual history, a set of 3D query points on an object of interest, and a language description of the intended goal, the model predicts the future 3D trajectory of each point. We introduce a full stack to study this task at scale: (1) MolmoMotion-1M is a large corpus of action-described, object-grounded 3D point trajectories annotated from 1.16M unconstrained videos; (2) PointMotionBench is a human-verified benchmark spanning 111 object categories and 61 motion types; and (3) MolmoMotion is a general motion forecasting model that supports both autoregressive coordinate prediction and flow-matching-based trajectory generation. MolmoMotion accurately predicts diverse motion patterns with different language instructions, and significantly outperforms existing motion prediction baselines on PointMotionBench. Finally, we show that the learned 3D motion prior transfers well to downstream applications: it improves training efficiency and generalization for robot manipulation, and its predicted trajectories provide effective motion guidance for generative models to synthesize videos with more realistic object motion. △ Less

Submitted 16 June, 2026; originally announced June 2026.

Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models

Authors: Mahtab Bigverdi, Linjie Li, Weikai Huang, Yiming Liu, Jaemin Cho, Tuhin Kundu, Chris Dongjoo Kim, Zelun Luo, Jieyu Zhang, Linda Shapiro, Ranjay Krishna

Abstract: Vision language models (VLMs) excel at many tasks but still struggle with spatial reasoning when critical information is not directly observable. Many such problems require imaginative perception: inferring what would be seen from an unseen viewpoint, tracing paths through occluded spaces, or integrating partial observations into a coherent spatial representation. We introduce Imaginative Percepti… ▽ More Vision language models (VLMs) excel at many tasks but still struggle with spatial reasoning when critical information is not directly observable. Many such problems require imaginative perception: inferring what would be seen from an unseen viewpoint, tracing paths through occluded spaces, or integrating partial observations into a coherent spatial representation. We introduce Imaginative Perception Tokens (IPT), intermediate perceptual representations that externalize what a VLM would perceive under alternative spatial configurations while remaining consistent with the observed input. To study this capability, we formulate three tasks, Perspective Taking (PET), Path Tracing (PT), and Multiview Counting (MVC), and construct datasets of approximately 20K examples with ground truth imaginations, answers, and evaluation benchmarks. Using the unified VLM BAGEL as the backbone, IPT supervision consistently improves spatial reasoning and often outperforms textual chain of thought training, even without generating images at inference time. On MVC, IPT improves accuracy by 3.4% and achieves competitive performance with strong closed-source models on PT. We further find that combining IPT and label-only supervision yields additional gains, whereas textual chain of thought can substantially degrade performance, suggesting a modality mismatch when spatial computation is forced through language. Overall, IPT provides a principled supervision signal for reasoning about unobserved spatial structure, improving generalization while producing interpretable intermediate representations. △ Less

Submitted 17 August, 2026; v1 submitted 2 June, 2026; originally announced June 2026.

JobBench: Aligning Agent Work With Human Will

Authors: Yuetai Li, Yichen Feng, Zhangchen Xu, Zixian Ma, Kaiyuan Zheng, Fengqing Jiang, Xinghua Sun, Rulin Shao, Zichen Chen, Yue Huang, Xinyang Han, Brian Lee, Kayla Xu, Shenglai Zeng, Hang Hua, Xiangliang Zhang, Basel Alomair, Ranjay Krishna, Luke Zettlemoyer, Pang Wei Koh, Bhaskar Ramasubramanian, Luyao Niu, Xiang Yue, Radha Poovendran

Abstract: Current benchmarks for occupational AI agents are scoped primarily by economic values, telling a replacement story. We introduce JobBench, which evaluates AI agents on the workflows that experts identify as high-priority for delegation, empowering humans based on their needs instead of replacing them with GDP value. JobBench covers 130 agentic tasks across 35 occupations. Each task is packaged as… ▽ More Current benchmarks for occupational AI agents are scoped primarily by economic values, telling a replacement story. We introduce JobBench, which evaluates AI agents on the workflows that experts identify as high-priority for delegation, empowering humans based on their needs instead of replacing them with GDP value. JobBench covers 130 agentic tasks across 35 occupations. Each task is packaged as a workspace of heterogeneous reference files, requiring the agent to reason through the cluttered information streams of real professional work. Outputs are graded by a fact-anchored chain of rubrics, averaging 35.6 binary criteria per task. We evaluate 36 models; the strongest, Claude Opus~4.7 under Claude Code, reaches only 45.9 %. We hope JobBench shifts the community's target labour-market effect from replacement to enhancement: building agents that do what humans actually want delegated, not only what is most economically valuable. △ Less

Submitted 25 May, 2026; originally announced May 2026.

Ablate-to-Validate: Are Vision-Language Models Really Using Continuous Thought Tokens?

Authors: Tianyi Zhang, Mahtab Bigverdi, Ranjay Krishna

Abstract: Vision-language models (VLMs) are increasingly augmented with continuous or latent non-textual tokens intended to support "visual thinking." Despite improved task accuracy, this alone does not show that models actually use these tokens for reasoning -- gains may arise from confounds such as added context length, special-token anchoring, or training-time regularization. We formalize a diagnostic pr… ▽ More Vision-language models (VLMs) are increasingly augmented with continuous or latent non-textual tokens intended to support "visual thinking." Despite improved task accuracy, this alone does not show that models actually use these tokens for reasoning -- gains may arise from confounds such as added context length, special-token anchoring, or training-time regularization. We formalize a diagnostic principle, Ablate-to-Validate, for testing whether latent-token content is genuinely utilized, and instantiate it as the Token Replacement Test (TRT), a standardized suite of content-replacement ablations. TRT holds the prompt, image, token budget, and decoding fixed while replacing intermediate tokens with zero, random, first-repeat, or oracle alternatives, isolating whether performance depends on token content or merely on token presence. As a controlled testbed, we study relative depth reasoning with LLaVA-13B and Qwen2.5-VL-3B, training models to predict and consume continuous or discrete depth spans across multiple frozen encoders (SigLIP2, CLIP, DINOv2) and token budgets. We additionally apply TRT to three off-the-shelf visual-thinking systems (Mirage, Mull-Tokens, CoVT) on BLINK, VSP, and CV-Bench. Across all settings, accuracy gains are a misleading proxy for latent-token reasoning: VLMs retain most improvement even when token content is corrupted or replaced, revealing a persistent gap between having a latent channel and using it as an information bottleneck. We recommend TRT as a standard diagnostic alongside accuracy for any method introducing continuous thought tokens. △ Less

Submitted 20 May, 2026; originally announced May 2026.

RefDecoder: Enhancing Visual Generation with Conditional Video Decoding

Authors: Xiang Fan, Yuheng Wang, Bohan Fang, Zhongzheng Ren, Ranjay Krishna

Abstract: Video generation powers a vast array of downstream applications. However, while the de facto standard, i.e., latent diffusion models, typically employ heavily conditioned denoising networks, their decoders often remain unconditional. We observe that this architectural asymmetry leads to significant loss of detail and inconsistency relative to the input image. To address this, we argue that the dec… ▽ More Video generation powers a vast array of downstream applications. However, while the de facto standard, i.e., latent diffusion models, typically employ heavily conditioned denoising networks, their decoders often remain unconditional. We observe that this architectural asymmetry leads to significant loss of detail and inconsistency relative to the input image. To address this, we argue that the decoder requires equal conditioning to preserve structural integrity. We introduce RefDecoder, a reference-conditioned video VAE decoder by injecting high-fidelity reference image signal directly into the decoding process via reference attention. Specifically, a lightweight image encoder maps the reference frame into the detail-rich high-dimensional tokens, which are co-processed with the denoised video latent tokens at each decoder up-sampling stage. We demonstrate consistent improvements across several distinct decoder backbones (e.g., Wan 2.1 and VideoVAE+), achieving up to +2.1dB PSNR over the unconditional baselines on the Inter4K, WebVid, and Large Motion reconstruction benchmarks. Notably, RefDecoder can be directly swapped into existing video generation systems without additional fine-tuning, and we report across-the-board improvements in subject consistency, background consistency, and overall quality scores on the VBench I2V benchmark. Beyond I2V, RefDecoder generalizes well to a wide range of visual generation tasks such as style transfer and video editing refinement. △ Less

Submitted 14 May, 2026; originally announced May 2026.

ScarfBench: A Benchmark for Cross-Framework Application Migration in Enterprise Java

Authors: Advait Pavuluri, Bridget McGinn, Ashita Saxena, George Safta, Srikanth Tamilselvam, Raju Pavuluri, Michele Merler, Baishakhi Ray, Rahul Krishna

Abstract: Java remains central to enterprise software, and many applications outlive their original architecture. Migrating them across frameworks is a behavior-preserving refactoring spanning build configuration, dependency injection, persistence, request handling, and deployment. Existing software-engineering benchmarks cover bug fixing, feature implementation, and language or version modernization, but l… ▽ More Java remains central to enterprise software, and many applications outlive their original architecture. Migrating them across frameworks is a behavior-preserving refactoring spanning build configuration, dependency injection, persistence, request handling, and deployment. Existing software-engineering benchmarks cover bug fixing, feature implementation, and language or version modernization, but leave cross-framework refactoring largely unmeasured. We introduce ScarfBench, a benchmark for behavior-preserving cross-framework refactoring of enterprise Java applications. It is built from expert-written implementation triples across Spring, Jakarta EE, and Quarkus: 34 applications (29 focused single-layer, 5 whole) yielding 102 variants (~151K lines across 1946 source and test files) and 204 directed refactoring tasks. Each task gives an agent a working source application and a target framework; the agent must synthesize a target implementation preserving the source behavior. Correctness is evaluated by an application-specific executable oracle: the candidate must compile, deploy in a containerized target runtime, and pass behavioral tests over the application's observable interface. We evaluate five state-of-the-art coding agents on ScarfBench. The strongest achieves only 15.3% aggregate test pass on focused-layer migrations and 12.2% on whole applications, and only one of the 204 tasks yields a fully behaviorally equivalent target. Difficulty is asymmetric across framework directions and architectural layers: Spring<->Quarkus is the most tractable pair, and Jakarta-targeted migrations are hardest. From LLM-as-a-judge and expert adjudication of failed-task traces, we derive a taxonomy of recurring failure categories spanning build, deploy, and test stages. We release the benchmark, harness, and agent traces at https://scarfbench.info. △ Less

Submitted 18 May, 2026; v1 submitted 7 May, 2026; originally announced May 2026.

Design Conductor 2.0: An agent builds a TurboQuant inference accelerator in 80 hours

Authors: The Verkor Team, Ravi Krishna, Suresh Krishna, David Chin

Abstract: Driven by a rapid co-evolution of both harness and underlying models, LLM agents are improving at a dizzying pace. In our prior work (performed in Dec. 2025), we introduced "Design Conductor" (or just "Conductor"), a system capable of building a 5-stage Linux-capable RISC-V CPU in 12 hours. In this work, we introduce an updated multi-agent harness powered by frontier models released in April 2026,… ▽ More Driven by a rapid co-evolution of both harness and underlying models, LLM agents are improving at a dizzying pace. In our prior work (performed in Dec. 2025), we introduced "Design Conductor" (or just "Conductor"), a system capable of building a 5-stage Linux-capable RISC-V CPU in 12 hours. In this work, we introduce an updated multi-agent harness powered by frontier models released in April 2026, which is able to handle 80x larger tasks, at higher quality, fully autonomously. Following a brief introduction, we examine 4 designs that the system produced autonomously, including "VerTQ", an LLM inference accelerator which hard-wires support for TurboQuant in a 240-cycle pipeline, starting from the TurboQuant arXiv paper. VerTQ includes heavy compute processing, with 5129 FP16/32 units; the design was mapped to an FPGA at 125 MHz and consumes 5.7 mm^2 in TSMC 16FF (8 attention pipes). We review the key new characteristics that enabled these results. Finally, we analyze Design Conductor's token usage and other empirical characteristics, including its limitations. △ Less

Submitted 6 May, 2026; originally announced May 2026.

MolmoAct2: Action Reasoning Models for Real-world Deployment

Authors: Haoquan Fang, Jiafei Duan, Donovan Clay, Sam Wang, Shuo Liu, Weikai Huang, Xiang Fan, Wei-Chuan Tsai, Shirui Chen, Yi Ru Wang, Shanli Xing, Jaemin Cho, Jae Sung Park, Ainaz Eftekhar, Peter Sushko, Karen Farley, Angad Wadhwa, Cole Harrison, Winson Han, Ying-Chun Lee, Eli VanderBilt, Rose Hendrix, Suveen Ellawela, Lucas Ngoo, Joyce Chai , et al. (4 additional authors not shown)

Abstract: Vision-Language-Action (VLA) models aim to provide a single generalist controller for robots, but today's systems fall short on the criteria that matter for real-world deployment. Frontier models are closed, open-weight alternatives are tied to expensive hardware, reasoning-augmented policies pay prohibitive latency for their grounding, and fine-tuned success rates remain below the threshold for d… ▽ More Vision-Language-Action (VLA) models aim to provide a single generalist controller for robots, but today's systems fall short on the criteria that matter for real-world deployment. Frontier models are closed, open-weight alternatives are tied to expensive hardware, reasoning-augmented policies pay prohibitive latency for their grounding, and fine-tuned success rates remain below the threshold for dependable use. We present MolmoAct2, a fully open action reasoning model built for practical deployment, advancing its predecessor along five axes. We introduce MolmoER, a VLM backbone specialized for spatial and embodied reasoning, trained on a 3.3M-sample corpus with a specialize-then-rehearse recipe. We release three new datasets spanning low-to-medium cost platforms, including MolmoAct2-BimanualYAM, 720 hours of teleoperated bimanual trajectories that constitute the largest open bimanual dataset to date, together with quality-filtered Franka (DROID) and SO100/101 subsets. We provide OpenFAST, an open-weight, open-data action tokenizer trained on millions of trajectories across five embodiments. We redesign the architecture to graft a flow-matching continuous-action expert onto a discrete-token VLM via per-layer KV-cache conditioning. Finally, we propose MolmoThink, an adaptive-depth reasoning variant that re-predicts depth tokens only for scene regions that change between timesteps, retaining geometric grounding at a fraction of prior latency. In the most extensive empirical study of any open VLA to date, spanning 7 simulation and real-world benchmarks, MolmoAct2 outperforms strong baselines including Pi-05, while MolmoER surpasses GPT-5 and Gemini Robotics ER-1.5 across 13 embodied-reasoning benchmarks. We release model weights, training code, and complete training data. Project page: https://allenai.org/blog/molmoact2 △ Less

Submitted 8 May, 2026; v1 submitted 4 May, 2026; originally announced May 2026.

Comments: 31 pages, project page: https://allenai.org/blog/molmoact2

VideoNet: A Large-Scale Dataset for Domain-Specific Action Recognition

Authors: Tanush Yadav, Mohammadreza Salehi, Jae Sung Park, Vivek Ramanujan, Hannaneh Hajishirzi, Yejin Choi, Ali Farhadi, Rohun Tripathi, Ranjay Krishna

Abstract: Videos are unique in their ability to capture actions which transcend multiple frames. Accordingly, for many years action recognition was the quintessential task for video understanding. Unfortunately, due to a lack of sufficiently diverse and challenging data, modern vision-language models (VLMs) are no longer evaluated on their action recognition capabilities. To revitalize action recognition in… ▽ More Videos are unique in their ability to capture actions which transcend multiple frames. Accordingly, for many years action recognition was the quintessential task for video understanding. Unfortunately, due to a lack of sufficiently diverse and challenging data, modern vision-language models (VLMs) are no longer evaluated on their action recognition capabilities. To revitalize action recognition in the era of VLMs, we advocate for a returned focus on domain-specific actions. To this end, we introduce VideoNet, a domain-specific action recognition benchmark covering 1,000 distinct actions from 37 domains. We begin with a multiple-choice evaluation setting, where the difference between closed and open models is stark: Gemini 3.1 Pro attains 69.9% accuracy while Qwen3-VL-8B gets a mere 45.0%. To understand why VLMs struggle on VideoNet, we relax the questions into a binary setting, where random chance is 50%. Still, Qwen achieves only 59.2% accuracy. Further relaxing the evaluation setup, we provide $k\in\{1,2,3\}$ in-context examples of the action. Some models excel in the few-shot setting, while others falter; Qwen improves $+7.0\%$, while Gemini declines $-4.8\%$. Notably, these gains fall short of the $+13.6\%$ improvement in non-expert humans when given few-shot examples. Finding that VLMs struggle to fully exploit in-context examples, we shift from test-time improvements to the training side. We collect the first large-scale training dataset for domain-specific actions, totaling nearly 500k video question-answer pairs. Fine-tuning a Molmo2-4B model on our data, we surpass all open-weight 8B models on the VideoNet benchmark. △ Less

Submitted 5 May, 2026; v1 submitted 4 May, 2026; originally announced May 2026.

Comments: CVPR 2026 Highlight. Website at https://tanu.sh/videonet

Posterior Augmented Flow Matching

Authors: George Stoica, Sayak Paul, Matthew Wallingford, Vivek Ramanujan, Abhay Nori, Winson Han, Ali Farhadi, Ranjay Krishna, Judy Hoffman

Abstract: Flow matching (FM) trains a time-dependent vector field that transports samples from a simple prior to a complex data distribution. However, for high-dimensional images, each training sample supervises only a single trajectory and intermediate point, yielding an extremely sparse and high-variance training signal. This under-constrained supervision can cause flow collapse, where the learned dynamic… ▽ More Flow matching (FM) trains a time-dependent vector field that transports samples from a simple prior to a complex data distribution. However, for high-dimensional images, each training sample supervises only a single trajectory and intermediate point, yielding an extremely sparse and high-variance training signal. This under-constrained supervision can cause flow collapse, where the learned dynamics memorize specific source-target pairings, mapping diverse inputs to overly similar outputs, failing to generalize. We introduce Posterior-Augmented Flow Matching (PAFM), a theoretically grounded generalization of FM that replaces single-target supervision with an expectation over an approximate posterior of valid target completions for a given intermediate state and condition. PAFM factorizes this intractable posterior into (i) the likelihood of the intermediate under a hypothesized endpoint and (ii) the prior probability of that endpoint under the condition, and uses an importance sampling scheme to construct a mixture over multiple candidate targets. We prove that PAFM yields an unbiased estimator of the original FM objective while substantially reducing gradient variance during training by aggregating information from many plausible continuation trajectories per intermediate. Finally, we show that PAFM improves over FM by up to 3.4 FID50K across different model scales (SiT-B/2 and SiT-XL/2), different architectures (SiT and MMDiT), and in both class and text conditioned benchmarks (ImageNet and CC12M), with a negligible increase in the compute overhead. Code: https://github.com/gstoica27/PAFM.git. △ Less

Submitted 1 May, 2026; originally announced May 2026.

You Only Judge Once: Multi-response Reward Modeling in a Single Forward Pass

Authors: Yinuo Yang, Zixian Ma, Manasi Ganti, Jieyu Zhang, Ranjay Krishna

Abstract: We present a discriminative multimodal reward model that scores all candidate responses in a single forward pass. Conventional discriminative reward models evaluate each response independently, requiring multiple forward passes, one for each potential response. Our approach concatenates multiple responses with separator tokens and applies cross-entropy over their scalar scores, enabling direct com… ▽ More We present a discriminative multimodal reward model that scores all candidate responses in a single forward pass. Conventional discriminative reward models evaluate each response independently, requiring multiple forward passes, one for each potential response. Our approach concatenates multiple responses with separator tokens and applies cross-entropy over their scalar scores, enabling direct comparative reasoning and efficient $N$-way preference learning. The multi-response design also yields up to $N\times$ wall-clock speedup and FLOPs reduction over conventional single-response scoring. To enable $N$-way reward evaluation beyond existing pairwise benchmarks, we construct two new benchmarks: (1) MR$^2$Bench-Image contains human-annotated rankings over responses from 8 diverse models; (2) MR$^2$Bench-Video is a large-scale video-based reward benchmark derived from 94K crowdsourced pairwise human judgments over video question-answering spanning 19 models, denoised via preference graph ensemble. Both benchmarks provide 4-response evaluation variants sampled from the full rankings. Built on a 4B vision-language backbone with LoRA fine-tuning and a lightweight MLP value head, our model achieves state-of-the-art results on six multimodal reward benchmarks, including MR$^2$Bench-Image, MR$^2$Bench-Video, and four other existing benchmarks. Our model outperforms existing larger generative and discriminative reward models. We further demonstrate that our reward model, when used in reinforcement learning with GRPO, produces improved policy models that maintain performance across standard multimodal benchmarks while substantially improving open-ended generation quality, outperforming a single-response discriminative reward model (RM) baseline by a large margin in both training stability and open-ended generation quality. △ Less

Submitted 15 April, 2026; v1 submitted 13 April, 2026; originally announced April 2026.

Comments: 9 pages, 4 figures

WildDet3D: Scaling Promptable 3D Detection in the Wild

Authors: Weikai Huang, Jieyu Zhang, Sijun Li, Taoyang Jia, Jiafei Duan, Yunqian Cheng, Jaemin Cho, Matthew Wallingford, Rustin Soraki, Chris Dongjoo Kim, Shuo Liu, Donovan Clay, Taira Anderson, Winson Han, Ali Farhadi, Bharath Hariharan, Zhongzheng Ren, Ranjay Krishna

Abstract: Understanding objects in 3D from a single image is a cornerstone of spatial intelligence. A key step toward this goal is monocular 3D object detection--recovering the extent, location, and orientation of objects from an input RGB image. To be practical in the open world, such a detector must generalize beyond closed-set categories, support diverse prompt modalities, and leverage geometric cues whe… ▽ More Understanding objects in 3D from a single image is a cornerstone of spatial intelligence. A key step toward this goal is monocular 3D object detection--recovering the extent, location, and orientation of objects from an input RGB image. To be practical in the open world, such a detector must generalize beyond closed-set categories, support diverse prompt modalities, and leverage geometric cues when available. Progress is hampered by two bottlenecks: existing methods are designed for a single prompt type and lack a mechanism to incorporate additional geometric cues, and current 3D datasets cover only narrow categories in controlled environments, limiting open-world transfer. In this work we address both gaps. First, we introduce WildDet3D, a unified geometry-aware architecture that natively accepts text, point, and box prompts and can incorporate auxiliary depth signals at inference time. Second, we present WildDet3D-Data, the largest open 3D detection dataset to date, constructed by generating candidate 3D boxes from existing 2D annotations and retaining only human-verified ones, yielding over 1M images across 13.5K categories in diverse real-world scenes. WildDet3D establishes a new state-of-the-art across multiple benchmarks and settings. In the open-world setting, it achieves 22.6/24.8 AP3D on our newly introduced WildDet3D-Bench with text and box prompts. On Omni3D, it reaches 34.2/36.4 AP3D with text and box prompts, respectively. In zero-shot evaluation, it achieves 40.3/48.9 ODS on Argoverse 2 and ScanNet. Notably, incorporating depth cues at inference time yields substantial additional gains (+20.7 AP on average across settings). △ Less

Submitted 17 April, 2026; v1 submitted 9 April, 2026; originally announced April 2026.

Comments: code: https://github.com/allenai/WildDet3D website: https://allenai.github.io/WildDet3D/

MolmoWeb: Open Visual Web Agent and Open Data for the Open Web

Authors: Tanmay Gupta, Piper Wolters, Zixian Ma, Peter Sushko, Rock Yuren Pang, Diego Llanes, Yue Yang, Taira Anderson, Boyuan Zheng, Zhongzheng Ren, Harsh Trivedi, Taylor Blanton, Caleb Ouellette, Winson Han, Ali Farhadi, Ranjay Krishna

Abstract: Web agents--autonomous systems that navigate and execute tasks on the web on behalf of users--have the potential to transform how people interact with the digital world. However, the most capable web agents today rely on proprietary models with undisclosed training data and recipes, limiting scientific understanding, reproducibility, and community-driven progress. We believe agents for the open… ▽ More Web agents--autonomous systems that navigate and execute tasks on the web on behalf of users--have the potential to transform how people interact with the digital world. However, the most capable web agents today rely on proprietary models with undisclosed training data and recipes, limiting scientific understanding, reproducibility, and community-driven progress. We believe agents for the open web should be built in the open. To this end, we introduce (1) MolmoWebMix, a large and diverse mixture of browser task demonstrations and web-GUI perception data and (2) MolmoWeb, a family of fully open multimodal web agents. Specifically, MolmoWebMix combines over 100K synthetic task trajectories from multiple complementary generation pipelines with 30K+ human demonstrations, atomic web-skill trajectories, and GUI perception data, including referring expression grounding and screenshot question answering. MolmoWeb agents operate as instruction-conditioned visual-language action policies: given a task instruction and a webpage screenshot, they predict the next browser action, requiring no access to HTML, accessibility trees, or specialized APIs. Available in 4B and 8B size, on browser-use benchmarks like WebVoyager, Online-Mind2Web, and DeepShop, MolmoWeb agents achieve state-of-the-art results outperforming similar scale open-weight-only models such as Fara-7B, UI-Tars-1.5-7B, and Holo1-7B. MolmoWeb-8B also surpasses set-of-marks (SoM) agents built on much larger closed frontier models like GPT-4o. We further demonstrate consistent gains through test-time scaling via parallel rollouts with best-of-N selection, achieving 94.7% and 60.5% pass@4 (compared to 78.2% and 35.3% pass@1) on WebVoyager and Online-Mind2Web respectively. We will release model checkpoints, training data, code, and a unified evaluation harness to enable reproducibility and accelerate open research on web agents. △ Less

Submitted 9 April, 2026; originally announced April 2026.

Comments: https://allenai.org/blog/molmoweb

MolmoPoint: Better Pointing for VLMs with Grounding Tokens

Authors: Christopher Clark, Yue Yang, Jae Sung Park, Zixian Ma, Jieyu Zhang, Rohun Tripathi, Mohammadreza Salehi, Sangho Lee, Taira Anderson, Winson Han, Ranjay Krishna

Abstract: Grounding has become a fundamental capability of vision-language models (VLMs). Most existing VLMs point by generating coordinates as part of their text output, which requires learning a complicated coordinate system and results in a high token count. Instead, we propose a more intuitive pointing mechanism that directly selects the visual tokens that contain the target concept. Our model generates… ▽ More Grounding has become a fundamental capability of vision-language models (VLMs). Most existing VLMs point by generating coordinates as part of their text output, which requires learning a complicated coordinate system and results in a high token count. Instead, we propose a more intuitive pointing mechanism that directly selects the visual tokens that contain the target concept. Our model generates a special pointing token that cross-attends to the input image or video tokens and selects the appropriate one. To make this model more fine-grained, we follow these pointing tokens with an additional special token that selects a fine-grained subpatch within the initially selected region, and then a third token that specifies a location within that subpatch. We further show that performance improves by generating points sequentially in a consistent order, encoding the relative position of the previously selected point, and including a special no-more-points class when selecting visual tokens. Using this method, we set a new state-of-the-art on image pointing (70.7% on PointBench), set a new state-of-the-art among fully open models on GUI pointing (61.1% on ScreenSpotPro), and improve video pointing (59.1% human preference win rate vs. a text coordinate baseline) and tracking (+6.3% gain on Molmo2Track). We additionally show that our method achieves much higher sample efficiency and discuss the qualitative differences that emerge from this design change. △ Less

Submitted 30 March, 2026; originally announced March 2026.

PerceptionComp: A Video Benchmark for Complex Perception-Centric Reasoning

Authors: Shaoxuan Li, Zhixuan Zhao, Hanze Deng, Zirun Ma, Shulin Tian, Zuyan Liu, Yushi Hu, Haoning Wu, Yuhao Dong, Benlin Liu, Ziwei Liu, Ranjay Krishna

Abstract: We introduce PerceptionComp, a manually annotated benchmark for complex, long-horizon, perception-centric video reasoning. PerceptionComp is designed so that no single moment is sufficient: answering each question requires multiple temporally separated pieces of visual evidence and compositional constraints under conjunctive and sequential logic, spanning perceptual subtasks such as objects, attri… ▽ More We introduce PerceptionComp, a manually annotated benchmark for complex, long-horizon, perception-centric video reasoning. PerceptionComp is designed so that no single moment is sufficient: answering each question requires multiple temporally separated pieces of visual evidence and compositional constraints under conjunctive and sequential logic, spanning perceptual subtasks such as objects, attributes, relations, locations, actions, and events, and requiring skills including semantic recognition, visual correspondence, temporal reasoning, and spatial reasoning. The benchmark contains 1,114 highly complex questions on 279 videos from diverse domains including city walk tours, indoor villa tours, video games, and extreme outdoor sports, with 100% manual annotation. Human studies show that PerceptionComp requires substantial test-time thinking and repeated perception steps: participants take much longer than on prior benchmarks, and accuracy drops to near chance (18.97%) when rewatching is disallowed. State-of-the-art MLLMs also perform substantially worse on PerceptionComp than on existing benchmarks: the best model in our evaluation, Gemini-3-Flash, reaches only 45.96% accuracy in the five-choice setting, while open-source models remain below 40%. These results suggest that perception-centric long-horizon video reasoning remains a major bottleneck, and we hope PerceptionComp will help drive progress in perceptual reasoning. △ Less

Submitted 27 March, 2026; originally announced March 2026.

Comments: Project Page: https://perceptioncomp.github.io

VFIG: Vectorizing Complex Figures in SVG with Vision-Language Models

Authors: Qijia He, Xunmei Liu, Hammaad Memon, Ziang Li, Zixian Ma, Jaemin Cho, Zhongzheng Ren, Daniel S Weld, Ranjay Krishna

Abstract: Scalable Vector Graphics (SVG) are essential for technical illustration and digital design, offering resolution independence and semantic editability. In practice, original vector files are frequently lost, leaving only rasterized versions (e.g., PNG, JPEG) that resist modification, while manual reconstruction is prohibitively expensive. Progress on automating raster-to-SVG conversion has been bot… ▽ More Scalable Vector Graphics (SVG) are essential for technical illustration and digital design, offering resolution independence and semantic editability. In practice, original vector files are frequently lost, leaving only rasterized versions (e.g., PNG, JPEG) that resist modification, while manual reconstruction is prohibitively expensive. Progress on automating raster-to-SVG conversion has been bottlenecked by two gaps: existing SVG datasets are dominated by icons and decorative graphics that lack the complexity of professional diagrams, and existing benchmarks rely on pixel- or embedding-level similarity that fails to capture structural correctness (e.g., broken connectivity, misplaced arrows). We close both gaps with paired contributions targeting diagram-centric figures (e.g., model architectures, flowcharts, schematics). For training, we introduce VFIG-Data, the largest figure-to-SVG dataset of its kind at 66K pairs, combining real paper figures converted via a describe-and-generate pipeline with programmatic diagrams that supply noise-free supervision over arrow styles, fonts, and geometry. For evaluation, we introduce VFIG-Bench, a structure-aware evaluation suite, paired with VFIG-Bench-OOD, an out-of-distribution set of figures manually curated from highly cited arXiv papers. Beyond pixel and embedding similarity, our protocol reports rubric-based VLM-Judge scores and Elo ratings from pairwise human preference evaluation. Built on these contributions, VFIG is a VLM family trained with a simple-to-complex SFT curriculum followed by RL with rendering-aware rewards. VFIG achieves state-of-the-art open-source performance, outperforming the best open-source VLM baseline by over 30%, and matches Claude Sonnet 4.6 on VFIG-BENCH: Gemini-Judge 78.2% vs. 76.7% and GPT-Judge 87.5% vs. 87.4%. It remains slightly behind the strongest proprietary models GPT-5.2 and Gemini-3. △ Less

Submitted 16 August, 2026; v1 submitted 25 March, 2026; originally announced March 2026.

Counting Circuits: Mechanistic Interpretability of Visual Reasoning in Large Vision-Language Models

Authors: Liwei Che, Zhiyu Xue, Yihao Quan, Benlin Liu, Zeru Shi, Michelle Hurst, Jacob Feldman, Ruixiang Tang, Ranjay Krishna, Vladimir Pavlovic

Abstract: Counting serves as a simple but powerful test of a Large Vision-Language Model's (LVLM's) reasoning; it forces the model to identify each individual object and then add them all up. In this study, we investigate how LVLMs implement counting using controlled synthetic and real-world benchmarks, combined with mechanistic analyses. Our results show that LVLMs display a human-like counting behavior, w… ▽ More Counting serves as a simple but powerful test of a Large Vision-Language Model's (LVLM's) reasoning; it forces the model to identify each individual object and then add them all up. In this study, we investigate how LVLMs implement counting using controlled synthetic and real-world benchmarks, combined with mechanistic analyses. Our results show that LVLMs display a human-like counting behavior, with precise performance on small numerosities and noisy estimation for larger quantities. We introduce two novel interpretability methods, Visual Activation Patching and HeadLens, and use them to uncover a structured "counting circuit" that is largely shared across a variety of visual reasoning tasks. Building on these insights, we propose a lightweight intervention strategy that exploits simple and abundantly available synthetic images to fine-tune arbitrary pretrained LVLMs exclusively on counting. Despite the narrow scope of this fine-tuning, the intervention not only enhances counting accuracy on in-distribution synthetic data, but also yields an average improvement of +8.36% on out-of-distribution counting benchmarks and an average gain of +1.54% on complex, general visual reasoning tasks for Qwen2.5-VL. These findings highlight the central, influential role of counting in visual reasoning and suggest a potential pathway for improving overall visual reasoning capabilities through targeted enhancement of counting mechanisms. △ Less

Submitted 19 March, 2026; originally announced March 2026.

Unified Spatio-Temporal Token Scoring for Efficient Video VLMs

Authors: Jianrui Zhang, Yue Yang, Rohun Tripathi, Winson Han, Ranjay Krishna, Christopher Clark, Yong Jae Lee, Sangho Lee

Abstract: Token pruning is essential for enhancing the computational efficiency of vision-language models (VLMs), particularly for video-based tasks where temporal redundancy is prevalent. Prior approaches typically prune tokens either (1) within the vision transformer (ViT) exclusively for unimodal perception tasks such as action recognition and object segmentation, without adapting to downstream vision-la… ▽ More Token pruning is essential for enhancing the computational efficiency of vision-language models (VLMs), particularly for video-based tasks where temporal redundancy is prevalent. Prior approaches typically prune tokens either (1) within the vision transformer (ViT) exclusively for unimodal perception tasks such as action recognition and object segmentation, without adapting to downstream vision-language tasks; or (2) only within the LLM while leaving the ViT output intact, often requiring complex text-conditioned token selection mechanisms. In this paper, we introduce Spatio-Temporal Token Scoring (STTS), a simple and lightweight module that prunes vision tokens across both the ViT and the LLM without text conditioning or token merging, and is fully compatible with end-to-end training. By learning how to score temporally via an auxiliary loss and spatially via LLM downstream gradients, aided by our efficient packing algorithm, STTS prunes 50% of vision tokens throughout the entire architecture, resulting in a 62% improvement in efficiency during both training and inference with only a 0.7% drop in average performance across 13 short and long video QA tasks. Efficiency gains increase with more sampled frames per video. Applying test-time scaling for long-video QA further yields performance gains of 0.5-1% compared to the baseline. Overall, STTS represents a novel, simple yet effective technique for unified, architecture-wide vision token pruning. △ Less

Submitted 18 March, 2026; originally announced March 2026.

MolmoB0T: Large-Scale Simulation Enables Zero-Shot Manipulation

Authors: Abhay Deshpande, Maya Guru, Rose Hendrix, Snehal Jauhri, Ainaz Eftekhar, Rohun Tripathi, Max Argus, Jordi Salvador, Haoquan Fang, Matthew Wallingford, Wilbert Pumacay, Yejin Kim, Quinn Pfeifer, Ying-Chun Lee, Piper Wolters, Omar Rayyan, Mingtong Zhang, Jiafei Duan, Karen Farley, Winson Han, Eli Vanderbilt, Dieter Fox, Ali Farhadi, Georgia Chalvatzaki, Dhruv Shah , et al. (1 additional authors not shown)

Abstract: A prevailing view in robot learning is that simulation alone is not enough; effective sim-to-real transfer is widely believed to require at least some real-world data collection or task-specific fine-tuning to bridge the gap between simulated and physical environments. We challenge that assumption. With sufficiently large-scale and diverse simulated synthetic training data, we show that zero-shot… ▽ More A prevailing view in robot learning is that simulation alone is not enough; effective sim-to-real transfer is widely believed to require at least some real-world data collection or task-specific fine-tuning to bridge the gap between simulated and physical environments. We challenge that assumption. With sufficiently large-scale and diverse simulated synthetic training data, we show that zero-shot transfer to the real world is not only possible, but effective for both static and mobile manipulation. We introduce MolmoBot-Engine, a fully open-source pipeline for procedural data generation across robots, tasks, and diverse simulated environments in MolmoSpaces. With it, we release MolmoBot-Data, a dataset of 1.8 million expert trajectories for articulated object manipulation and pick-and-place tasks. We train three policy classes: MolmoBot, a Molmo2-based multi-frame vision-language model with a flow-matching action head; MolmoBot-Pi0, which replicates the $π_0$ architecture to enable direct comparison; and MolmoBot-SPOC, a lightweight policy suitable for edge deployment and amenable to RL fine-tuning. We evaluate on two robotic platforms: the Franka FR3 for tabletop manipulation tasks and the Rainbow Robotics RB-Y1 mobile manipulator for door opening, drawer manipulation, cabinet interaction, and mobile pick-and-place. Without any real-world fine-tuning, our policies achieve zero-shot transfer to unseen objects and environments. On tabletop pick-and-place, MolmoBot achieves a success rate of 79.2% in real world evaluations across 4 settings, outperforming $π_{0.5}$ at 39.2%. Our results demonstrate that procedural environment generation combined with diverse articulated assets can produce robust manipulation policies that generalize broadly to the real world. Technical website: https://allenai.github.io/MolmoBot △ Less

Submitted 26 March, 2026; v1 submitted 17 March, 2026; originally announced March 2026.

Electrodermal Activity as a Unimodal Signal for Aerobic Exercise Detection in Wearable Sensors

Authors: Rena Mira Krishna, Ramya Sankar, Shadi Ghiasi

Abstract: Electrodermal Activity (EDA) is a non-invasive physiological signal widely available in wearable devices and reflects sympathetic nervous system (SNS) activation. Prior multi-modal studies have demonstrated robust performance in distinguishing stress and exercise states when EDA is combined with complementary signals such as heart rate and accelerometry. However, the ability of EDA to independentl… ▽ More Electrodermal Activity (EDA) is a non-invasive physiological signal widely available in wearable devices and reflects sympathetic nervous system (SNS) activation. Prior multi-modal studies have demonstrated robust performance in distinguishing stress and exercise states when EDA is combined with complementary signals such as heart rate and accelerometry. However, the ability of EDA to independently distinguish sustained aerobic exercise from low-arousal states under subject-independent evaluation remains insufficiently characterized. This study investigates whether features derived exclusively from EDA can reliably differentiate rest from sustained aerobic exercise. Using a publicly available dataset collected from thirty healthy individuals, EDA features were evaluated using benchmark machine learning models with leave-one-subject-out (LOSO) validation. Across models, EDA-only classifiers achieved moderate subject-independent performance, with phasic temporal dynamics and event timing contributing to class separation. Rather than proposing EDA as a replacement for multimodal sensing, this work provides a conservative benchmark of the discriminative power of EDA alone and clarifies its role as a unimodal input for wearable activity-state inference. △ Less

Submitted 16 March, 2026; originally announced March 2026.

Journal ref: Conference paper: Proceedings of the 52nd Annual Northeast Bioengineering Conference (NEBEC 2026), 04/17/2026, Philadelphia, USA

vla-eval: A Unified Evaluation Harness for Vision-Language-Action Models

Authors: Suhwan Choi, Yunsung Lee, Yubeen Park, Chris Dongjoo Kim, Ranjay Krishna, Dieter Fox, Youngjae Yu

Abstract: Vision-Language-Action (VLA) models are increasingly evaluated across multiple simulation benchmarks, yet adding each benchmark to an evaluation pipeline requires resolving incompatible dependencies, matching underspecified evaluation protocols, and reverse-engineering undocumented preprocessing. This burden scales with the number of models and benchmarks, making comprehensive evaluation impractic… ▽ More Vision-Language-Action (VLA) models are increasingly evaluated across multiple simulation benchmarks, yet adding each benchmark to an evaluation pipeline requires resolving incompatible dependencies, matching underspecified evaluation protocols, and reverse-engineering undocumented preprocessing. This burden scales with the number of models and benchmarks, making comprehensive evaluation impractical for most teams. We present vla-eval, an open-source evaluation harness that eliminates this per-benchmark cost by decoupling model inference from benchmark execution through a WebSocket+msgpack protocol with Docker-based environment isolation. Models integrate once by implementing a single predict() method; benchmarks integrate once via a four-method interface; the full cross-evaluation matrix works automatically. The framework supports 14 simulation benchmarks and six model servers. Parallel evaluation via episode sharding and batch inference achieves up to 47x wall-clock speedup, completing 2,000 LIBERO episodes in ~18 minutes. To validate the framework, we reproduce published scores across six VLA codebases and three benchmarks, documenting previously undocumented pitfalls. We additionally release a VLA leaderboard aggregating 657 published results across 17 benchmarks. Framework, evaluation configs, and all reproduction results are publicly available at https://github.com/allenai/vla-evaluation-harness and https://allenai.github.io/vla-evaluation-harness/leaderboard. △ Less

Submitted 17 April, 2026; v1 submitted 14 March, 2026; originally announced March 2026.

Video-Based Reward Modeling for Computer-Use Agents

Authors: Linxin Song, Jieyu Zhang, Huanxin Sheng, Taiwei Shi, Gupta Rahul, Yang Liu, Ranjay Krishna, Jian Kang, Jieyu Zhao

Abstract: Computer-using agents (CUAs) are becoming increasingly capable; however, it remains difficult to scale evaluation of whether a trajectory truly fulfills a user instruction. In this work, we study reward modeling from execution video: a sequence of keyframes from an agent trajectory that is independent of the agent's internal reasoning or actions. Although video-execution modeling is method-agnosti… ▽ More Computer-using agents (CUAs) are becoming increasingly capable; however, it remains difficult to scale evaluation of whether a trajectory truly fulfills a user instruction. In this work, we study reward modeling from execution video: a sequence of keyframes from an agent trajectory that is independent of the agent's internal reasoning or actions. Although video-execution modeling is method-agnostic, it presents key challenges, including highly redundant layouts and subtle, localized cues that determine success. We introduce Execution Video Reward 53k (ExeVR-53k), a dataset of 53k high-quality video--task--reward triplets. We further propose adversarial instruction translation to synthesize negative samples with step-level annotations. To enable learning from long, high-resolution execution videos, we design spatiotemporal token pruning, which removes homogeneous regions and persistent tokens while preserving decisive UI changes. Building on these components, we fine-tune an Execution Video Reward Model (ExeVRM) that takes only a user instruction and a video-execution sequence to predict task success. Our ExeVRM 8B achieves 84.7% accuracy and 87.7% recall on video-execution assessment, outperforming strong proprietary models such as GPT-5.2 and Gemini-3 Pro across Ubuntu, macOS, Windows, and Android, while providing more precise temporal attribution. These results show that video-execution reward modeling can serve as a scalable, model-agnostic evaluator for CUAs. △ Less

Submitted 10 March, 2026; originally announced March 2026.

Design Conductor: An agent autonomously builds a 1.5 GHz Linux-capable RISC-V CPU

Authors: The Verkor Team, Ravi Krishna, Suresh Krishna, David Chin

Abstract: Design Conductor (DC) is an autonomous agent which applies the capabilities of frontier models to build semiconductors end-to-end -- that is, from concept to verified, tape-out ready GDSII (layout CAD file). In 12 hours and fully autonomously, DC was able to build several micro-architecture variations of a complete RISC-V CPU (which we dub VerCore) that meet timing at 1.48 GHz (rv32i-zmmul; using… ▽ More Design Conductor (DC) is an autonomous agent which applies the capabilities of frontier models to build semiconductors end-to-end -- that is, from concept to verified, tape-out ready GDSII (layout CAD file). In 12 hours and fully autonomously, DC was able to build several micro-architecture variations of a complete RISC-V CPU (which we dub VerCore) that meet timing at 1.48 GHz (rv32i-zmmul; using the ASAP7 PDK), starting from a 219-word requirements document. The VerCore achieves a CoreMark score of 3261. For historical context, this is roughly equivalent to an Intel Celeron SU2300 from mid-2011 (which ran at 1.2 GHz). To our knowledge, this is the first time an autonomous agent has built a complete, working CPU from spec to GDSII. This report is organized as follows. We first review DC's design and its key components. We then describe the methodology that DC followed to build VerCore -- including RTL implementation, testbench implementation, frontend debugging, optimization to achieve timing closure, and interacting with backend tools. We review the key characteristics of the resulting VerCore. Finally, we highlight how frontier models could improve to better enable this application, and our lessons learned as to how chips will be built in the future enabled by the capabilities of systems like DC. △ Less

Submitted 6 February, 2026; originally announced March 2026.

When Rubrics Fail: Error Enumeration as Reward in Reference-Free RL Post-Training for Virtual Try-On

Authors: Wisdom Ikezogwo, Mehmet Saygin Seyfioglu, Ranjay Krishna, Karim Bouyarmane

Abstract: Reinforcement learning with verifiable rewards (RLVR) and Rubrics as Rewards (RaR) have driven strong gains in domains with clear correctness signals and even in subjective domains by synthesizing evaluation criteria from ideal reference answers. But many real-world tasks admit multiple valid outputs and lack the single ideal answer that rubric generation depends on. We identify this reference-fre… ▽ More Reinforcement learning with verifiable rewards (RLVR) and Rubrics as Rewards (RaR) have driven strong gains in domains with clear correctness signals and even in subjective domains by synthesizing evaluation criteria from ideal reference answers. But many real-world tasks admit multiple valid outputs and lack the single ideal answer that rubric generation depends on. We identify this reference-free setting as a gap in current post-training methods and propose Implicit Error Counting (IEC) to fill it. Instead of checking what a response gets right against a rubric, IEC enumerates what it gets wrong, applying severity-weighted scores across task-relevant axes and converting them into calibrated per-aspect rewards. We show that naïve explicit enumeration is too noisy for stable optimization, and that two design choices: implicit score emission and group calibration are necessary to make error counting a reliable reward. As a case study, we validate IEC on virtual try-on (VTO), a domain that is simultaneously too constrained for holistic scoring and too permissive for rubric-based evaluation: subtle garment errors are unacceptable, yet many output variations are correct. We introduce Cascaded Error Counting (CEC) as an evaluation metric, which tracks human preferences well (60% top-1 vs. 30% others), and curate Mismatch-DressCode (MDressBench), a benchmark with maximal attribute mismatch to stress-test reward designs. On MDressBench, IEC outperforms RaR across all metrics (CEC: 5.31 vs. 5.60 on flat references; 5.20 vs. 5.53 on non-flat). On VITON-HD and DressCode, IEC matches or surpasses six baselines on 6 of 8 perceptual metrics. These results suggest that when ideal answers are unavailable, counting errors provide a stronger signal than constructing rubrics. △ Less

Submitted 6 July, 2026; v1 submitted 5 March, 2026; originally announced March 2026.

Synthetic Visual Genome 2: Extracting Large-scale Spatio-Temporal Scene Graphs from Videos

Authors: Ziqi Gao, Jieyu Zhang, Wisdom Oluchi Ikezogwo, Jae Sung Park, Tario G. You, Daniel Ogbu, Chenhao Zheng, Weikai Huang, Yinuo Yang, Winson Han, Quan Kong, Rajat Saini, Ranjay Krishna

Abstract: We introduce Synthetic Visual Genome 2 (SVG2), a large-scale panoptic video scene graph dataset. SVG2 contains over 636K videos with 6.6M objects, 52.0M attributes, and 6.7M relations, providing an order-of-magnitude increase in scale and diversity over prior spatio-temporal scene graph datasets. To create SVG2, we design a fully automated pipeline that combines multi-scale panoptic segmentation,… ▽ More We introduce Synthetic Visual Genome 2 (SVG2), a large-scale panoptic video scene graph dataset. SVG2 contains over 636K videos with 6.6M objects, 52.0M attributes, and 6.7M relations, providing an order-of-magnitude increase in scale and diversity over prior spatio-temporal scene graph datasets. To create SVG2, we design a fully automated pipeline that combines multi-scale panoptic segmentation, online-offline trajectory tracking with automatic new-object discovery, per-trajectory semantic parsing, and GPT-5-based spatio-temporal relation inference. Human verification of SVG2 annotation accuracy confirms its reliability (objects: 93.8%, attributes: 88.3%, relations: 85.4%). Building on this resource, we train TRaSER, a video scene graph generation model. TRaSER augments VLMs with a trajectory-aligned token arrangement mechanism and new modules: an object-trajectory resampler and a temporal-window resampler to convert raw videos and panoptic trajectories into compact spatio-temporal scene graphs in a single forward pass. The temporal-window resampler binds visual tokens to short trajectory segments to preserve local motion and temporal semantics, while the object-trajectory resampler aggregates entire trajectories to maintain global context for objects. On the PVSG, VIPSeg, VidOR and SVG2 test datasets, TRaSER improves relation detection by +15 to 20%, object prediction by +30 to 40% over the strongest open-source baselines and by +13% over GPT-5, and attribute prediction by +15%. When TRaSER's generated scene graphs are sent to a VLM for video question answering, it delivers a +1.5 to 4.6% absolute accuracy gain over using video only or video augmented with Qwen2.5-VL's generated scene graphs, demonstrating the utility of explicit spatio-temporal scene graphs as an intermediate representation. △ Less

Submitted 6 March, 2026; v1 submitted 26 February, 2026; originally announced February 2026.

Scale Can't Overcome Pragmatics: The Impact of Reporting Bias on Vision-Language Reasoning

Authors: Amita Kamath, Jack Hessel, Khyathi Chandu, Jena D. Hwang, Kai-Wei Chang, Ranjay Krishna

Abstract: The lack of reasoning capabilities in Vision-Language Models (VLMs) has remained at the forefront of research discourse. We posit that this behavior stems from a reporting bias in their training data. That is, how people communicate about visual content by default omits tacit information needed to supervise some types of reasoning; e.g., "at the game today!" is a more likely caption than "a photo… ▽ More The lack of reasoning capabilities in Vision-Language Models (VLMs) has remained at the forefront of research discourse. We posit that this behavior stems from a reporting bias in their training data. That is, how people communicate about visual content by default omits tacit information needed to supervise some types of reasoning; e.g., "at the game today!" is a more likely caption than "a photo of 37 people standing behind a field". We investigate the data underlying the popular VLMs OpenCLIP, LLaVA-1.5 and Molmo through the lens of theories from pragmatics, and find that reporting bias results in insufficient representation of four reasoning skills (spatial, temporal, negation, and counting), despite the corpora being of web-scale, and/or synthetically generated. With a set of curated benchmarks, we demonstrate that: (i) VLMs perform poorly on the aforementioned types of reasoning suppressed in the training data by reporting bias; (ii) contrary to popular belief, scaling data size, model size, and to multiple languages does not result in emergence of these skills by default; but, promisingly, (iii) incorporating annotations specifically collected to obtain tacit information is effective. Our findings highlight the need for more intentional training data curation methods, rather than counting on scale for emergence of reasoning capabilities. △ Less

Submitted 26 February, 2026; originally announced February 2026.

Comments: TACL 2026

TrajTok: Learning Trajectory Tokens enables better Video Understanding

Authors: Chenhao Zheng, Jieyu Zhang, Jianing Zhang, Weikai Huang, Ashutosh Kumar, Quan Kong, Oncel Tuzel, Chun-Liang Li, Ranjay Krishna

Abstract: Tokenization in video models, typically through patchification, generates an excessive and redundant number of tokens. This severely limits video efficiency and scalability. While recent trajectory-based tokenizers offer a promising solution by decoupling video duration from token count, they rely on complex external segmentation and tracking pipelines that are slow and task-agnostic. We propose T… ▽ More Tokenization in video models, typically through patchification, generates an excessive and redundant number of tokens. This severely limits video efficiency and scalability. While recent trajectory-based tokenizers offer a promising solution by decoupling video duration from token count, they rely on complex external segmentation and tracking pipelines that are slow and task-agnostic. We propose TrajTok, an end-to-end video tokenizer module that is fully integrated and co-trained with video models for a downstream objective, dynamically adapting its token granularity to semantic complexity, independent of video duration. TrajTok contains a unified segmenter that performs implicit clustering over pixels in both space and time to directly produce object trajectories in a single forward pass. By prioritizing downstream adaptability over pixel-perfect segmentation fidelity, TrajTok is lightweight and efficient, yet empirically improves video understanding performance. With TrajTok, we implement a video CLIP model trained from scratch (TrajViT2). It achieves the best accuracy at scale across both classification and retrieval benchmarks, while maintaining efficiency comparable to the best token-merging methods. TrajTok also proves to be a versatile component beyond its role as a tokenizer. We show that it can be seamlessly integrated as either a probing head for pretrained visual features (TrajAdapter) or an alignment connector in vision-language models (TrajVLM) with especially strong performance in long-video reasoning. △ Less

Submitted 3 June, 2026; v1 submitted 26 February, 2026; originally announced February 2026.

Comments: CVPR 2026

TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics

Authors: Shirui Chen, Cole Harrison, Ying-Chun Lee, Angela Jin Yang, Zhongzheng Ren, Lillian J. Ratliff, Jiafei Duan, Dieter Fox, Ranjay Krishna

Abstract: General-purpose robot learning requires dense, instruction-conditioned feedback that can distinguish meaningful task progress from stalled, failed, or partially completed behavior. Yet obtaining such feedback at scale remains difficult, since existing approaches often rely on manual progress annotations, task-specific demonstrations, or reward models trained on curated robot datasets. We introduce… ▽ More General-purpose robot learning requires dense, instruction-conditioned feedback that can distinguish meaningful task progress from stalled, failed, or partially completed behavior. Yet obtaining such feedback at scale remains difficult, since existing approaches often rely on manual progress annotations, task-specific demonstrations, or reward models trained on curated robot datasets. We introduce TOPReward, a training-free progress reward method that probes pretrained Video-Language Models (VLMs) through their internal token probabilities rather than asking them to generate numerical progress values. Given a video prefix and a language instruction, TOPReward measures the model's likelihood that the instructed task has been completed, converting latent video-language understanding into a dense reward signal without task-specific reward-model training or manually annotated progress labels. We evaluate TOPReward on ManiRewardBench, our real-world manipulation benchmark spanning 130 unique tasks and four robot platforms, as well as on Open X-Embodiment datasets. Across these settings, TOPReward substantially outperforms prior training-free VLM reward methods on open-source models and is competitive with a trained reward-model baseline on progress-estimation metrics, while requiring no reward-model training. Additional analyses show that the reward is sensitive to the specified instruction and is not explained by time index alone. Finally, TOPReward supports downstream applications including success detection and offline reward-weighted behavior cloning. △ Less

Submitted 22 July, 2026; v1 submitted 22 February, 2026; originally announced February 2026.

CapNav: Benchmarking Vision Language Models on Capability-conditioned Indoor Navigation

Authors: Xia Su, Ruiqi Chen, Benlin Liu, Jingwei Ma, Zonglin Di, Ranjay Krishna, Jon Froehlich

Abstract: Vision-Language Models (VLMs) have shown remarkable progress in Vision-Language Navigation (VLN), offering new possibilities for navigation decision-making that could benefit both robotic platforms and human users. However, real-world navigation is inherently conditioned by the agent's mobility constraints. For example, a sweeping robot cannot traverse stairs, while a quadruped can. We introduce C… ▽ More Vision-Language Models (VLMs) have shown remarkable progress in Vision-Language Navigation (VLN), offering new possibilities for navigation decision-making that could benefit both robotic platforms and human users. However, real-world navigation is inherently conditioned by the agent's mobility constraints. For example, a sweeping robot cannot traverse stairs, while a quadruped can. We introduce Capability-Conditioned Navigation (CapNav), a benchmark designed to evaluate how well VLMs can navigate complex indoor spaces given an agent's specific physical and operational capabilities. CapNav defines five representative human and robot agents, each described with physical dimensions, mobility capabilities, and environmental interaction abilities. CapNav provides 45 real-world indoor scenes, 473 navigation tasks, and 2365 QA pairs to test if VLMs can traverse indoor environments based on agent capabilities. We evaluate 13 modern VLMs and find that current VLM's navigation performance drops sharply as mobility constraints tighten, and that even state-of-the-art models struggle with obstacle types that require reasoning on spatial dimensions. We conclude by discussing the implications for capability-aware navigation and the opportunities for advancing embodied spatial reasoning in future VLMs. The benchmark is available at https://github.com/makeabilitylab/CapNav △ Less

Submitted 20 February, 2026; originally announced February 2026.

MolmoSpaces: A Large-Scale Open Ecosystem for Robot Navigation and Manipulation

Authors: Yejin Kim, Wilbert Pumacay, Omar Rayyan, Max Argus, Winson Han, Eli VanderBilt, Jordi Salvador, Abhay Deshpande, Rose Hendrix, Snehal Jauhri, Shuo Liu, Nur Muhammad Mahi Shafiullah, Maya Guru, Ainaz Eftekhar, Karen Farley, Donovan Clay, Jiafei Duan, Arjun Guru, Piper Wolters, Alvaro Herrasti, Ying-Chun Lee, Georgia Chalvatzaki, Yuchen Cui, Ali Farhadi, Dieter Fox , et al. (1 additional authors not shown)

Abstract: Deploying robots at scale demands robustness to the long tail of everyday situations. The countless variations in scene layout, object geometry, and task specifications that characterize real environments are vast and underrepresented in existing robot benchmarks. Measuring this level of generalization requires infrastructure at a scale and diversity that physical evaluation alone cannot provide.… ▽ More Deploying robots at scale demands robustness to the long tail of everyday situations. The countless variations in scene layout, object geometry, and task specifications that characterize real environments are vast and underrepresented in existing robot benchmarks. Measuring this level of generalization requires infrastructure at a scale and diversity that physical evaluation alone cannot provide. We introduce MolmoSpaces, a fully open ecosystem to support large-scale benchmarking of robot policies. MolmoSpaces consists of over 230k diverse indoor environments, ranging from handcrafted household scenes to procedurally generated multiroom houses, populated with 130k richly annotated object assets, including 48k manipulable objects with 42M stable grasps. Crucially, these environments are simulator-agnostic, supporting popular options such as MuJoCo, Isaac, and ManiSkill. The ecosystem supports the full spectrum of embodied tasks: static and mobile manipulation, navigation, and multiroom long-horizon tasks requiring coordinated perception, planning, and interaction across entire indoor environments. We also design MolmoSpaces-Bench, a benchmark suite of 8 tasks in which robots interact with our diverse scenes and richly annotated objects. Our experiments show MolmoSpaces-Bench exhibits strong sim-to-real correlation (R = 0.96, \r{ho} = 0.98), confirm newer and stronger zero-shot policies outperform earlier versions in our benchmarks, and identify key sensitivities to prompt phrasing, initial joint positions, and camera occlusion. Through MolmoSpaces and its open-source assets and tooling, we provide a foundation for scalable data generation, policy training, and benchmark creation for robot learning research. △ Less

Submitted 18 February, 2026; v1 submitted 11 February, 2026; originally announced February 2026.

Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative Reasoning

Authors: Yalcin Tur, Jalal Naghiyev, Haoquan Fang, Wei-Chuan Tsai, Jiafei Duan, Dieter Fox, Ranjay Krishna

Abstract: Current Vision-Language-Action (VLA) models rely on fixed computational depth, expending the same amount of compute on simple adjustments and complex multi-step manipulation. While Chain-of-Thought (CoT) prompting enables variable computation, it scales memory linearly and is ill-suited for continuous action spaces. We introduce Recurrent-Depth VLA (RD-VLA), an architecture that achieves computati… ▽ More Current Vision-Language-Action (VLA) models rely on fixed computational depth, expending the same amount of compute on simple adjustments and complex multi-step manipulation. While Chain-of-Thought (CoT) prompting enables variable computation, it scales memory linearly and is ill-suited for continuous action spaces. We introduce Recurrent-Depth VLA (RD-VLA), an architecture that achieves computational adaptivity via latent iterative refinement rather than explicit token generation. RD-VLA employs a recurrent, weight-tied action head that supports arbitrary inference depth with a constant memory footprint. The model is trained using truncated backpropagation through time (TBPTT) to efficiently supervise the refinement process. At inference, RD-VLA dynamically allocates compute using an adaptive stopping criterion based on latent convergence. Experiments on challenging manipulation tasks show that recurrent depth is critical: tasks that fail entirely (0 percent success) with single-iteration inference exceed 90 percent success with four iterations, while simpler tasks saturate rapidly. RD-VLA provides a scalable path to test-time compute in robotics, replacing token-based reasoning with latent reasoning to achieve constant memory usage and up to 80x inference speedup over prior reasoning-based VLA models. Project page: https://rd-vla.github.io/ △ Less

Submitted 8 February, 2026; originally announced February 2026.

Comments: 11 Pages, Project page:https://rd-vla.github.io/

Theory of Space: Can Foundation Models Construct Spatial Beliefs through Active Exploration?

Authors: Pingyue Zhang, Zihan Huang, Yue Wang, Jieyu Zhang, Letian Xue, Zihan Wang, Qineng Wang, Keshigeyan Chandrasegaran, Ruohan Zhang, Yejin Choi, Ranjay Krishna, Jiajun Wu, Li Fei-Fei, Manling Li

Abstract: Spatial embodied intelligence requires agents to act to acquire information under partial observability. While multimodal foundation models excel at passive perception, their capacity for active, self-directed exploration remains understudied. We propose Theory of Space, defined as an agent's ability to actively acquire information through self-directed, active exploration and to construct, revise… ▽ More Spatial embodied intelligence requires agents to act to acquire information under partial observability. While multimodal foundation models excel at passive perception, their capacity for active, self-directed exploration remains understudied. We propose Theory of Space, defined as an agent's ability to actively acquire information through self-directed, active exploration and to construct, revise, and exploit a spatial belief from sequential, partial observations. We evaluate this through a benchmark where the goal is curiosity-driven exploration to build an accurate cognitive map. A key innovation is spatial belief probing, which prompts models to reveal their internal spatial representations at each step. Our evaluation of state-of-the-art models reveals several critical bottlenecks. First, we identify an Active-Passive Gap, where performance drops significantly when agents must autonomously gather information. Second, we find high inefficiency, as models explore unsystematically compared to program-based proxies. Through belief probing, we diagnose that while perception is an initial bottleneck, global beliefs suffer from instability that causes spatial knowledge to degrade over time. Finally, using a false belief paradigm, we uncover Belief Inertia, where agents fail to update obsolete priors with new evidence. This issue is present in text-based agents but is particularly severe in vision-based models. Our findings suggest that current foundation models struggle to maintain coherent, revisable spatial beliefs during active exploration. △ Less

Submitted 4 February, 2026; originally announced February 2026.

Comments: published at iclr 2026

VLS: Steering Pretrained Robot Policies via Vision-Language Models

Authors: Shuo Liu, Ishneet Sukhvinder Singh, Yiqing Xu, Jiafei Duan, Ranjay Krishna

Abstract: Why do pretrained diffusion or flow-matching policies fail when the same task is performed near an obstacle, on a shifted support surface, or amid mild clutter? Such failures rarely reflect missing motor skills; instead, they expose a limitation of imitation learning under train-test shifts, where action generation is tightly coupled to training-specific spatial configurations and task specificati… ▽ More Why do pretrained diffusion or flow-matching policies fail when the same task is performed near an obstacle, on a shifted support surface, or amid mild clutter? Such failures rarely reflect missing motor skills; instead, they expose a limitation of imitation learning under train-test shifts, where action generation is tightly coupled to training-specific spatial configurations and task specifications. Retraining or fine-tuning to address these failures is costly and conceptually misaligned, as the required behaviors already exist but cannot be selectively adapted at test time. We propose Vision-Language Steering (VLS), a training-free framework for inference-time adaptation of frozen generative robot policies. VLS treats adaptation as an inference-time control problem, steering the sampling process of a pretrained diffusion or flow-matching policy in response to out-of-distribution observation-language inputs without modifying policy parameters. By leveraging vision-language models to synthesize trajectory-differentiable reward functions, VLS guides denoising toward action trajectories that satisfy test-time spatial and task requirements. Across simulation and real-world evaluations, VLS consistently outperforms prior steering methods, achieving a 31% improvement on CALVIN and a 13% gain on LIBERO-PRO. Real-world deployment on a Franka robot further demonstrates robust inference-time adaptation under test-time spatial and semantic shifts. Project page: https://vision-language-steering.github.io/webpage/ △ Less

Submitted 3 February, 2026; originally announced February 2026.

Comments: 11 Pages, Project page: https://vision-language-steering.github.io/webpage/

Multi-task Code LLMs: Data Mix or Model Merge?

Authors: Mingzhi Zhu, Boris Sobolev, Rahul Krishna, Raju Pavuluri, Stacy Patterson, Michele Merler

Abstract: Recent research advocates deploying smaller, specialized code LLMs in agentic frameworks alongside frontier models, sparking interest in efficient strategies for multi-task learning that balance performance, constraints, and costs. We compare two approaches for creating small, multi-task code LLMs: data mixing versus model merging. We conduct extensive experiments across two model families (Qwen C… ▽ More Recent research advocates deploying smaller, specialized code LLMs in agentic frameworks alongside frontier models, sparking interest in efficient strategies for multi-task learning that balance performance, constraints, and costs. We compare two approaches for creating small, multi-task code LLMs: data mixing versus model merging. We conduct extensive experiments across two model families (Qwen Coder and DeepSeek Coder) at two scales (2B and 7B parameters), fine-tuning them for code generation and code summarization tasks. Our evaluation on HumanEval, MBPP, and CodeXGlue benchmarks reveals that model merging achieves the best overall performance at larger scale across model families, retaining 96% of specialized model performance on code generation tasks while maintaining summarization capabilities. Notably, merged models can even surpass individually fine-tuned models, with our best configuration of Qwen Coder 2.5 7B model achieving 92.7% Pass@1 on HumanEval compared to 90.9% for its task-specific fine-tuned equivalent. At a smaller scale we find instead data mixing to be a preferred strategy. We further introduce a weight analysis technique to understand how different tasks affect model parameters and their implications for merging strategies. The results suggest that careful merging and mixing strategies can effectively combine task-specific capabilities without significant performance degradation, making them ideal for resource-constrained deployment scenarios. △ Less

Submitted 28 January, 2026; originally announced January 2026.

Usage, Effects and Requirements for AI Coding Assistants in the Enterprise: An Empirical Study

Authors: Maja Vukovic, Rangeet Pan, Tin Kam Ho, Rahul Krishna, Raju Pavuluri, Michele Merler

Abstract: The rise of large language models (LLMs) has accelerated the development of automated techniques and tools for supporting various software engineering tasks, e.g., program understanding, code generation, software testing, and program repair. As CodeLLMs are being employed toward automating these tasks, one question that arises, especially in enterprise settings, is whether these coding assistants… ▽ More The rise of large language models (LLMs) has accelerated the development of automated techniques and tools for supporting various software engineering tasks, e.g., program understanding, code generation, software testing, and program repair. As CodeLLMs are being employed toward automating these tasks, one question that arises, especially in enterprise settings, is whether these coding assistants and the code LLMs that power them are ready for real-world projects and enterprise use cases, and how do they impact the existing software engineering process and user experience. In this paper we survey 57 developers from different domains and with varying software engineering skill about their experience with AI coding assistants and CodeLLMs. We also reviewed 35 user surveys on the usage, experience and expectations of professionals and students using AI coding assistants and CodeLLMs. Based on our study findings and analysis of existing surveys, we discuss the requirements for AI-powered coding assistants. △ Less

Submitted 27 January, 2026; originally announced January 2026.

Comments: To appear in the 3rd International Workshop on Large Language Models For Code, co-located at ICSE, Rio de Janeiro, Brazil, 2026

AdaReasoner: Dynamic Tool Orchestration for Iterative Visual Reasoning

Authors: Mingyang Song, Haoyu Sun, Jiawei Gu, Linjie Li, Luxin Xu, Ranjay Krishna, Yu Cheng

Abstract: When humans face problems beyond their immediate capabilities, they rely on tools, providing a promising paradigm for improving visual reasoning in multimodal large language models (MLLMs). Effective reasoning, therefore, hinges on knowing which tools to use, when to invoke them, and how to compose them over multiple steps, even when faced with new tools or new tasks. We introduce \textbf{AdaReaso… ▽ More When humans face problems beyond their immediate capabilities, they rely on tools, providing a promising paradigm for improving visual reasoning in multimodal large language models (MLLMs). Effective reasoning, therefore, hinges on knowing which tools to use, when to invoke them, and how to compose them over multiple steps, even when faced with new tools or new tasks. We introduce \textbf{AdaReasoner}, a family of multimodal models that learn tool use as a general reasoning skill rather than as tool-specific or explicitly supervised behavior. AdaReasoner is enabled by (i) a scalable data curation pipeline exposing models to long-horizon, multi-step tool interactions; (ii) Tool-GRPO, a reinforcement learning algorithm that optimizes tool selection and sequencing based on end-task success; and (iii) an adaptive learning mechanism that dynamically regulates tool usage. Together, these components allow models to infer tool utility from task context and intermediate outcomes, enabling coordination of multiple tools and generalization to unseen tools. Empirically, AdaReasoner exhibits strong tool-adaptive and generalization behaviors: it autonomously adopts beneficial tools, suppresses irrelevant ones, and adjusts tool usage frequency based on task demands, despite never being explicitly trained to do so. These capabilities translate into state-of-the-art performance across challenging benchmarks, improving the 7B base model by +24.9\% on average and surpassing strong proprietary systems such as GPT-5 on multiple tasks, including VSP and Jigsaw. △ Less

Submitted 28 January, 2026; v1 submitted 26 January, 2026; originally announced January 2026.

Comments: 28 pages, 10 figures and 13 tables

Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding

Authors: Christopher Clark, Jieyu Zhang, Zixian Ma, Jae Sung Park, Mohammadreza Salehi, Rohun Tripathi, Sangho Lee, Zhongzheng Ren, Chris Dongjoo Kim, Yinuo Yang, Vincent Shao, Yue Yang, Weikai Huang, Ziqi Gao, Taira Anderson, Jianrui Zhang, Jitesh Jain, George Stoica, Winson Han, Ali Farhadi, Ranjay Krishna

Abstract: Today's strongest video-language models (VLMs) remain proprietary. The strongest open-weight models either rely on synthetic data from proprietary VLMs, effectively distilling from them, or do not disclose their training data or recipe. As a result, the open-source community lacks the foundations needed to improve on the state-of-the-art video (and image) language models. Crucially, many downstrea… ▽ More Today's strongest video-language models (VLMs) remain proprietary. The strongest open-weight models either rely on synthetic data from proprietary VLMs, effectively distilling from them, or do not disclose their training data or recipe. As a result, the open-source community lacks the foundations needed to improve on the state-of-the-art video (and image) language models. Crucially, many downstream applications require more than just high-level video understanding; they require grounding -- either by pointing or by tracking in pixels. Even proprietary models lack this capability. We present Molmo2, a new family of VLMs that are state-of-the-art among open-source models and demonstrate exceptional new capabilities in point-driven grounding in single image, multi-image, and video tasks. Our key contribution is a collection of 7 new video datasets and 2 multi-image datasets, including a dataset of highly detailed video captions for pre-training, a free-form video Q&A dataset for fine-tuning, a new object tracking dataset with complex queries, and an innovative new video pointing dataset, all collected without the use of closed VLMs. We also present a training recipe for this data utilizing an efficient packing and message-tree encoding scheme, and show bi-directional attention on vision tokens and a novel token-weight strategy improves performance. Our best-in-class 8B model outperforms others in the class of open weight and data models on short videos, counting, and captioning, and is competitive on long-videos. On video-grounding Molmo2 significantly outperforms existing open-weight models like Qwen3-VL (35.5 vs 29.6 accuracy on video counting) and surpasses proprietary models like Gemini 3 Pro on some tasks (38.4 vs 20.0 F1 on video pointing and 56.2 vs 41.1 J&F on video tracking). △ Less

Submitted 2 April, 2026; v1 submitted 15 January, 2026; originally announced January 2026.

Comments: Updated first authors

PRAXIS: Integrating Program Analysis with Observability for Root-Cause Analysis

Authors: Shengkun Cui, Rahul Krishna, Saurabh Jha, Ravishankar K. Iyer

Abstract: Unresolved production cloud incidents cost an average of over $2M per hour. This paper introduces PRAXIS, an orchestrator that manages and deploys an agentic workflow for diagnosing code- and configuration-caused cloud incidents. PRAXIS employs an LLM-driven structured traversal over two types of graph: (1) a service dependency graph (SDG) that captures microservice-level dependencies; and (2) a h… ▽ More Unresolved production cloud incidents cost an average of over $2M per hour. This paper introduces PRAXIS, an orchestrator that manages and deploys an agentic workflow for diagnosing code- and configuration-caused cloud incidents. PRAXIS employs an LLM-driven structured traversal over two types of graph: (1) a service dependency graph (SDG) that captures microservice-level dependencies; and (2) a hammock-block program dependence graph (PDG) that captures code-level dependencies for each microservice. Compared to state-of-the-art ReAct baselines, PRAXIS improves RCA accuracy by up to 6.3x while reducing token consumption by 5.3x. PRAXIS is demonstrated on a set of 30 comprehensive real-world incidents that is being compiled into an RCA benchmark. △ Less

Submitted 28 April, 2026; v1 submitted 26 December, 2025; originally announced December 2025.

Comments: 15 pages. Accepted to appear in The 56th Annual IEEE/IFIP International Conference on Dependable Systems and Networks

GenEval 2: Addressing Benchmark Drift in Text-to-Image Evaluation

Authors: Amita Kamath, Kai-Wei Chang, Ranjay Krishna, Luke Zettlemoyer, Yushi Hu, Marjan Ghazvininejad

Abstract: Automating Text-to-Image (T2I) model evaluation is challenging; a judge model must be used to score correctness, and test prompts must be selected to be challenging for current T2I models but not the judge. We argue that satisfying these constraints can lead to benchmark drift over time, where the static benchmark judges fail to keep up with newer model capabilities. We show that benchmark drift i… ▽ More Automating Text-to-Image (T2I) model evaluation is challenging; a judge model must be used to score correctness, and test prompts must be selected to be challenging for current T2I models but not the judge. We argue that satisfying these constraints can lead to benchmark drift over time, where the static benchmark judges fail to keep up with newer model capabilities. We show that benchmark drift is a significant problem for GenEval, one of the most popular T2I benchmarks. Although GenEval was well-aligned with human judgment at the time of its release, it has drifted far from human judgment over time -- resulting in an absolute error of as much as 17.7% for current models. This level of drift strongly suggests that GenEval has been saturated for some time, as we verify via a large-scale human study. To help fill this benchmarking gap, we introduce a new benchmark, GenEval 2, with improved coverage of primitive visual concepts and higher degrees of compositionality, which we show is more challenging for current models. We also introduce Soft-TIFA, an evaluation method for GenEval 2 that combines judgments for visual primitives, which we show is more well-aligned with human judgment and argue is less likely to drift from human-alignment over time (as compared to more holistic judges such as VQAScore). Although we hope GenEval 2 will provide a strong benchmark for many years, avoiding benchmark drift is far from guaranteed and our work, more generally, highlights the importance of continual audits and improvement for T2I and related automated model evaluation benchmarks. △ Less

Submitted 18 December, 2025; originally announced December 2025.

Structure From Tracking: Distilling Structure-Preserving Motion for Video Generation

Authors: Yang Fei, George Stoica, Jingyuan Liu, Qifeng Chen, Ranjay Krishna, Xiaojuan Wang, Benlin Liu

Abstract: Reality is a dance between rigid constraints and deformable structures. For video models, that means generating motion that preserves fidelity as well as structure. Despite progress in diffusion models, producing realistic structure-preserving motion remains challenging, especially for articulated and deformable objects such as humans and animals. Scaling training data alone, so far, has failed to… ▽ More Reality is a dance between rigid constraints and deformable structures. For video models, that means generating motion that preserves fidelity as well as structure. Despite progress in diffusion models, producing realistic structure-preserving motion remains challenging, especially for articulated and deformable objects such as humans and animals. Scaling training data alone, so far, has failed to resolve physically implausible transitions. Existing approaches rely on conditioning with noisy motion representations, such as optical flow or skeletons extracted using an external imperfect model. To address these challenges, we introduce an algorithm to distill structure-preserving motion priors from an autoregressive video tracking model (SAM2) into a bidirectional video diffusion model (CogVideoX). With our method, we train SAM2VideoX, which contains two innovations: (1) a bidirectional feature fusion module that extracts global structure-preserving motion priors from a recurrent model like SAM2; (2) a Local Gram Flow loss that aligns how local features move together. Experiments on VBench and in human studies show that SAM2VideoX delivers consistent gains (+2.60\% on VBench, 21-22\% lower FVD, and 71.4\% human preference) over prior baselines. Specifically, on VBench, we achieve 95.51\%, surpassing REPA (92.91\%) by 2.60\%, and reduce FVD to 360.57, a 21.20\% and 22.46\% improvement over REPA- and LoRA-finetuning, respectively. The project website can be found at https://sam2videox.github.io/ . △ Less

Submitted 12 December, 2025; originally announced December 2025.

Comments: Project Website: https://sam2videox.github.io/

Mull-Tokens: Modality-Agnostic Latent Thinking

Authors: Arijit Ray, Ahmed Abdelkader, Chengzhi Mao, Bryan A. Plummer, Kate Saenko, Ranjay Krishna, Leonidas Guibas, Wen-Sheng Chu

Abstract: Reasoning goes beyond language; the real world requires reasoning about space, time, affordances, and much more that words alone cannot convey. Existing multimodal models exploring the potential of reasoning with images are brittle and do not scale. They rely on calling specialist tools, costly generation of images, or handcrafted reasoning data to switch between text and image thoughts. Instead,… ▽ More Reasoning goes beyond language; the real world requires reasoning about space, time, affordances, and much more that words alone cannot convey. Existing multimodal models exploring the potential of reasoning with images are brittle and do not scale. They rely on calling specialist tools, costly generation of images, or handcrafted reasoning data to switch between text and image thoughts. Instead, we offer a simpler alternative -- Mull-Tokens -- modality-agnostic latent tokens pre-trained to hold intermediate information in either image or text modalities to let the model think free-form towards the correct answer. We investigate best practices to train Mull-Tokens inspired by latent reasoning frameworks. We first train Mull-Tokens using supervision from interleaved text-image traces, and then fine-tune without any supervision by only using the final answers. Across four challenging spatial reasoning benchmarks involving tasks such as solving puzzles and taking different perspectives, we demonstrate that Mull-Tokens improve upon several baselines utilizing text-only reasoning or interleaved image-text reasoning, achieving a +3% average improvement and up to +16% on a puzzle solving reasoning-heavy split compared to our strongest baseline. Adding to conversations around challenges in grounding textual and visual reasoning, Mull-Tokens offers a simple solution to abstractly think in multiple modalities. △ Less

Submitted 30 April, 2026; v1 submitted 11 December, 2025; originally announced December 2025.

Comments: Project webpage: https://arijitray.com/multimodal_thinking/, Accepted to CVPR 2026 (Findings Track)

OmniView: An All-Seeing Diffusion Model for 3D and 4D View Synthesis

Authors: Xiang Fan, Sharath Girish, Vivek Ramanujan, Chaoyang Wang, Ashkan Mirzaei, Petr Sushko, Aliaksandr Siarohin, Sergey Tulyakov, Ranjay Krishna

Abstract: Prior approaches injecting camera control into diffusion models have focused on specific subsets of 4D consistency tasks: novel view synthesis, text-to-video with camera control, image-to-video, amongst others. Therefore, these fragmented approaches are trained on disjoint slices of available 3D/4D data. We introduce OmniView, a unified framework that generalizes across a wide range of 4D consiste… ▽ More Prior approaches injecting camera control into diffusion models have focused on specific subsets of 4D consistency tasks: novel view synthesis, text-to-video with camera control, image-to-video, amongst others. Therefore, these fragmented approaches are trained on disjoint slices of available 3D/4D data. We introduce OmniView, a unified framework that generalizes across a wide range of 4D consistency tasks. Our method separately represents space, time, and view conditions, enabling flexible combinations of these inputs. For example, OmniView can synthesize novel views from static, dynamic, and multiview inputs, extrapolate trajectories forward and backward in time, and create videos from text or image prompts with full camera control. OmniView is competitive with task-specific models across diverse benchmarks and metrics, improving image quality scores among camera-conditioned diffusion models by up to 33\% in multiview NVS LLFF dataset, 60\% in dynamic NVS Neural 3D Video benchmark, 20\% in static camera control on RE-10K, and reducing camera trajectory errors by 4x in text-conditioned video generation. With strong generalizability in one model, OmniView demonstrates the feasibility of a generalist 4D video model. Project page is available at https://snap-research.github.io/OmniView/ △ Less

Submitted 21 January, 2026; v1 submitted 11 December, 2025; originally announced December 2025.

Comments: Project page: https://snap-research.github.io/OmniView/

Agile Deliberation: Concept Deliberation for Subjective Visual Classification

Authors: Leijie Wang, Otilia Stretcu, Wei Qiao, Thomas Denby, Krishnamurthy Viswanathan, Enming Luo, Chun-Ta Lu, Tushar Dogra, Ranjay Krishna, Ariel Fuxman

Abstract: From content moderation to content curation, applications requiring vision classifiers for visual concepts are rapidly expanding. Existing human-in-the-loop approaches typically assume users begin with a clear, stable concept understanding to be able to provide high-quality supervision. In reality, users often start with a vague idea and must iteratively refine it through "concept deliberation", a… ▽ More From content moderation to content curation, applications requiring vision classifiers for visual concepts are rapidly expanding. Existing human-in-the-loop approaches typically assume users begin with a clear, stable concept understanding to be able to provide high-quality supervision. In reality, users often start with a vague idea and must iteratively refine it through "concept deliberation", a practice we uncovered through structured interviews with content moderation experts. We operationalize the common strategies in deliberation used by real content moderators into a human-in-the-loop framework called "Agile Deliberation" that explicitly supports evolving and subjective concepts. The system supports users in defining the concept for themselves by exposing them to borderline cases. The system does this with two deliberation stages: (1) concept scoping, which decomposes the initial concept into a structured hierarchy of sub-concepts, and (2) concept iteration, which surfaces semantically borderline examples for user reflection and feedback to iteratively align an image classifier with the user's evolving intent. Since concept deliberation is inherently subjective and interactive, we painstakingly evaluate the framework through 18 user sessions, each 1.5h long, rather than standard benchmarking datasets. We find that Agile Deliberation achieves 7.5% higher F1 scores than automated decomposition baselines and more than 3% higher than manual deliberation, while participants reported clearer conceptual understanding and lower cognitive effort. △ Less

Submitted 3 April, 2026; v1 submitted 11 December, 2025; originally announced December 2025.

Journal ref: CVPR 2026

Process-Centric Analysis of Agentic Software Systems

Authors: Shuyang Liu, Yang Chen, Rahul Krishna, Saurabh Sinha, Jatin Ganhotra, Reyhan Jabbarvand

Abstract: Agentic systems are modern software systems: they consist of orchestrated modules, expose interfaces, and are deployed in software pipelines. Unlike conventional programs, their execution, i.e., trajectories, is inherently stochastic and adaptive to the problems they solve. Evaluation of such systems is often outcome-centric. This narrow focus overlooks detailed insights, failing to explain how ag… ▽ More Agentic systems are modern software systems: they consist of orchestrated modules, expose interfaces, and are deployed in software pipelines. Unlike conventional programs, their execution, i.e., trajectories, is inherently stochastic and adaptive to the problems they solve. Evaluation of such systems is often outcome-centric. This narrow focus overlooks detailed insights, failing to explain how agents reason, plan, act, or change their strategies. Inspired by the structured representation of conventional software systems as graphs, we introduce Graphectory to systematically encode the temporal and semantic relations in such systems. Using Graphectory, we automatically analyze 4000 trajectories of two dominant agentic programming workflows, SWE-agent and OpenHands, with four backbone Large Language Models (LLMs), attempting to resolve SWE-bench issues. Our automated analyses (completed within four minutes) reveal that: (1) agents using richer prompts or stronger LLMs exhibit more complex Graphectory, reflecting deeper exploration, broader context gathering, and more thorough validation; (2) agents' strategies vary with problem difficulty and the underlying LLM - for resolved issues, strategies often follow coherent localization-patching-validation steps, while unresolved ones exhibit chaotic or backtracking behaviors; and (3) even successful agentic systems often display inefficient processes. We also implement a novel technique for real-time construction and analysis of Graphectory and Langutory during agent execution to flag trajectory issues. Upon detecting such issues, the technique notifies the agent with a diagnostic message and, when applicable, rolls back the trajectory. Experiments show that online monitoring and interventions improve resolution rates by 6.9%-23.5% across models for problematic instances, while significantly shortening trajectories with near-zero overhead. △ Less

Submitted 10 April, 2026; v1 submitted 1 December, 2025; originally announced December 2025.

OlmoEarth: Stable Latent Image Modeling for Multimodal Earth Observation

Authors: Henry Herzog, Favyen Bastani, Yawen Zhang, Gabriel Tseng, Joseph Redmon, Hadrien Sablon, Ryan Park, Jacob Morrison, Alexandra Buraczynski, Karen Farley, Joshua Hansen, Andrew Howe, Patrick Alan Johnson, Mark Otterlee, Ted Schmitt, Hunter Pitelka, Stephen Daspit, Rachel Ratner, Christopher Wilhelm, Sebastian Wood, Mike Jacobi, Hannah Kerner, Evan Shelhamer, Ali Farhadi, Ranjay Krishna , et al. (1 additional authors not shown)

Abstract: Earth observation data presents a unique challenge: it is spatial like images, sequential like video or text, and highly multimodal. We present OlmoEarth: a multimodal, spatio-temporal foundation model that employs a novel self-supervised learning formulation, masking strategy, and loss all designed for the Earth observation domain. OlmoEarth achieves state-of-the-art performance compared to 12 ot… ▽ More Earth observation data presents a unique challenge: it is spatial like images, sequential like video or text, and highly multimodal. We present OlmoEarth: a multimodal, spatio-temporal foundation model that employs a novel self-supervised learning formulation, masking strategy, and loss all designed for the Earth observation domain. OlmoEarth achieves state-of-the-art performance compared to 12 other foundation models across a variety of research benchmarks and real-world tasks from external partners. When evaluating embeddings OlmoEarth achieves the best performance on 15 out of 24 tasks, and with full fine-tuning it is the best on 19 of 29 tasks. We deploy OlmoEarth as the backbone of an end-to-end platform for data collection, labeling, training, and inference of Earth observation models. The OlmoEarth Platform puts frontier foundation models and powerful data management tools into the hands of non-profits and NGOs working to solve the world's biggest problems. OlmoEarth source code, training data, and pre-trained weights are available at $\href{https://github.com/allenai/olmoearth_pretrain}{\text{https://github.com/allenai/olmoearth_pretrain}}$. △ Less

Submitted 17 November, 2025; originally announced November 2025.

SAINT: Service-level Integration Test Generation with Program Analysis and LLM-based Agents

Authors: Rangeet Pan, Raju Pavuluri, Ruikai Huang, Rahul Krishna, Tyler Stennett, Alessandro Orso, Saurabh SInha

Abstract: Enterprise applications are typically tested at multiple levels, with service-level testing playing an important role in validating application functionality. Existing service-level testing tools, especially for RESTful APIs, often employ fuzzing and/or depend on OpenAPI specifications which are not readily available in real-world enterprise codebases. Moreover, these tools are limited in their ab… ▽ More Enterprise applications are typically tested at multiple levels, with service-level testing playing an important role in validating application functionality. Existing service-level testing tools, especially for RESTful APIs, often employ fuzzing and/or depend on OpenAPI specifications which are not readily available in real-world enterprise codebases. Moreover, these tools are limited in their ability to generate functional tests that effectively exercise meaningful scenarios. In this work, we present SAINT, a novel white-box testing approach for service-level testing of enterprise Java applications. SAINT combines static analysis, large language models (LLMs), and LLM-based agents to automatically generate endpoint and scenario-based tests. The approach builds two key models: an endpoint model, capturing syntactic and semantic information about service endpoints, and an operation dependency graph, capturing inter-endpoint ordering constraints. SAINT then employs LLM-based agents to generate tests. Endpoint-focused tests aim to maximize code and database interaction coverage. Scenario-based tests are synthesized by extracting application use cases from code and refining them into executable tests via planning, action, and reflection phases of the agentic loop. We evaluated SAINT on eight Java applications, including a proprietary enterprise application. Our results illustrate the effectiveness of SAINT in coverage, fault detection, and scenario generation. Moreover, a developer survey provides strong endorsement of the scenario-based tests generated by SAINT. Overall, our work shows that combining static analysis with agentic LLM workflows enables more effective, functional, and developer-aligned service-level test generation. △ Less

Submitted 24 May, 2026; v1 submitted 17 November, 2025; originally announced November 2025.

Comments: Accepted at ICSE'26

DiScoFormer: Plug-In Density and Score Estimation with Transformers

Authors: Vasily Ilin, Peter Sushko, Ranjay Krishna

Abstract: Estimating probability density and its score from samples remains a core problem in generative modeling, Bayesian inference, and kinetic theory. Existing methods are bifurcated: classical kernel density estimators (KDE) generalize across distributions but suffer from the curse of dimensionality, while modern neural score models achieve high precision but require retraining for every target distrib… ▽ More Estimating probability density and its score from samples remains a core problem in generative modeling, Bayesian inference, and kinetic theory. Existing methods are bifurcated: classical kernel density estimators (KDE) generalize across distributions but suffer from the curse of dimensionality, while modern neural score models achieve high precision but require retraining for every target distribution. We introduce DiScoFormer (Density and Score Transformer), a ``train-once, infer-anywhere" equivariant Transformer that maps i.i.d. samples to both density values and score vectors, generalizing across distributions and sample sizes. Analytically, we prove that self-attention can recover normalized KDE, establishing it as a functional generalization of kernel methods; empirically, individual attention heads learn multi-scale, kernel-like behaviors. The model converges faster and achieves higher precision than KDE for density estimation, and provides a high-fidelity plug-in score oracle for score-debiased KDE, Fisher information computation, and Fokker-Planck-type PDEs. △ Less

Submitted 2 June, 2026; v1 submitted 8 November, 2025; originally announced November 2025.

Comments: Accepted in ICML 2026 (oral)

MSC Class: 68T07; 62G07 ACM Class: I.2.6; G.3

SIMS-V: Simulated Instruction-Tuning for Spatial Video Understanding

Authors: Ellis Brown, Arijit Ray, Ranjay Krishna, Ross Girshick, Rob Fergus, Saining Xie

Abstract: Despite impressive high-level video comprehension, multimodal language models struggle with spatial reasoning across time and space. While current spatial training approaches rely on real-world video data, obtaining diverse footage with precise spatial annotations remains a bottleneck. To alleviate this bottleneck, we present SIMS-V -- a systematic data-generation framework that leverages the priv… ▽ More Despite impressive high-level video comprehension, multimodal language models struggle with spatial reasoning across time and space. While current spatial training approaches rely on real-world video data, obtaining diverse footage with precise spatial annotations remains a bottleneck. To alleviate this bottleneck, we present SIMS-V -- a systematic data-generation framework that leverages the privileged information of 3D simulators to create spatially-rich video training data for multimodal language models. Using this framework, we investigate which properties of simulated data drive effective real-world transfer through systematic ablations of question types, mixes, and scales. We identify a minimal set of three question categories (metric measurement, perspective-dependent reasoning, and temporal tracking) that prove most effective for developing transferable spatial intelligence, outperforming comprehensive coverage despite using fewer question types. These insights enable highly efficient training: our 7B-parameter video LLM fine-tuned on just 25K simulated examples outperforms the larger 72B baseline and achieves competitive performance with proprietary models on rigorous real-world spatial reasoning benchmarks. Our approach demonstrates robust generalization, maintaining performance on general video understanding while showing substantial improvements on embodied and real-world spatial tasks. △ Less

Submitted 13 November, 2025; v1 submitted 6 November, 2025; originally announced November 2025.

Comments: Project page: https://ellisbrown.github.io/sims-v