Tag
HDR is a unified framework integrating hierarchical latents into causal video generation for multi-step visual reasoning, achieving better reasoning consistency, lower latency, and strong data efficiency compared to baselines.
UniVR introduces VR-GRPO, a reinforcement learning paradigm for unified visual reasoning, learning complex reasoning and physical dynamics from pure visual demonstrations, achieving up to 25% improvement on the VR-X benchmark.
Muse Spark 1.1 outperforms GPT-5.6 Sol and Gemini 3.1 on Radiology's Last Exam 2.0, a new visual reasoning benchmark for autonomous AI diagnosis in healthcare, though it still lags behind Fable and human radiologists.
Moondream 3.1 is a vision language model with mixture-of-experts architecture (9B total parameters, 2B active), delivering state-of-the-art visual reasoning, detection, pointing, and captioning, deployable locally via the Photon inference engine or through the Moondream Cloud API.
This paper shows that state-of-the-art vision-language models fail at length generalization in visual reasoning due to 'global shortcuts', and demonstrates that combining local foveated perception with recurrence enables robust out-of-distribution generalization.
MentalThink introduces a visual-symbolic reasoning paradigm for multimodal LLMs that uses SVG code as intermediate visual representations for multi-turn reasoning. The two-stage training with SFT and RL achieves strong performance on spatial reasoning benchmarks.
PixelEyes proposes a multi-turn visual reasoning agent that decouples perception and reasoning using mask-guided search and semantic-region breadth-first search, introducing a new benchmark (Pinpoint-Bench) and dataset (PixelEyes-6K) to improve localization in visual evidence seeking.
V-Zero is a novel label-free framework for fine-grained visual reasoning that uses contrastive evidence gating and on-policy distillation to improve performance without annotated answer labels, achieving faster training than traditional methods.
VeriEvol is a novel framework for scaling reinforcement learning in visual mathematical reasoning by ensuring reliable reward labels through a two-axis approach separating prompt difficulty from answer reliability, using evolutionary operators and hypothesis-testing verification. It achieves significant accuracy gains on a five-benchmark visual-math suite.
This paper introduces visually grounded thinking, a method for vision-language models to interleave natural-language reasoning with explicit visual evidence grounding using points or boxes. A scalable synthesis pipeline and grounding-aware reinforcement learning improve reasoning accuracy, enabling a 4B model to match or surpass a 27B model on spatial and counting benchmarks.
Visual Para-Thinker++ proposes a single-policy multi-agent framework for visual reasoning that uses role-conditioned agents (Main, Worker, Summary) and dedicated training methods to reduce hallucinations and improve efficiency, outperforming baselines on hallucination-sensitive benchmarks.
This paper identifies that failures in visual reasoning often stem from breakdowns in dynamic cross-modal coordination between visual and textual evidence during chain-of-thought generation. It introduces DyCo-RL, a reinforcement learning framework that rewards effective cross-modal coordination, leading to improved reasoning performance.
Introduces Efficient Operator Search (EOS), a unified differentiable framework that generalizes token reduction methods (pruning, merging, pooling, adaptive reweighting) into a shared operator space, automatically searching for optimal operator compositions under budget constraints. The method achieves competitive results across benchmarks and reveals consistent operator patterns.
VAMPS is a new benchmark of 1,168 multimodal bilingual math problems designed to evaluate whether LLMs can benefit from constructing and reasoning over graphs/visualizations. Key finding: direct analytical solving surprisingly outperforms tool-enabled visual solving even on problems where plotting is a natural strategy.
TRON introduces a scalable online environment for visual reasoning reinforcement learning that generates unlimited diverse training instances with verifiable answers, showing consistent performance improvements across multiple multimodal benchmarks.
VisualThink-VLA introduces a visual intermediate reasoning framework for vision-language-action policies that preserves spatial precision and dramatically reduces latency compared to text-based reasoning, achieving sub-second inference and state-of-the-art success rates on robot manipulation benchmarks.
This paper proposes MARS, a mono-anchored multi-source reasoning framework that uses dynamic anchors to quantify information gain and regulate modality interactions during reinforcement learning with verifiable rewards, achieving 3.2% and 4.9% performance gains on GRPO and DAPO across diverse datasets.
ETCHR is a novel image editing approach that decouples visual reasoning from image generation, using a two-stage training process (Reasoning Imitation and Reasoning Enhancement) to improve multimodal language model performance across five visual reasoning tasks. It achieves consistent gains of 4-5% Pass@1 on models like Qwen3-VL-8B, Gemini-3.1-Flash-Lite, and Kimi K2.5.
ATLAS presents a visual reasoning framework that combines agentic operations and latent representations using functional tokens, enabling efficient training via next-token prediction and reinforcement learning while avoiding intermediate image generation.
This paper introduces RIS, a framework for spatial-semantic grounded latent visual reasoning in Multimodal Large Language Models to overcome information bottlenecks. It proposes anchoring latent tokens to spatial and semantic evidence, showing improvements on benchmarks like V* and HRBench.