multimodal-reasoning 开源项目推荐（55 个）

tongjingqi /Thinking-with-Video

我们介绍“视频思维”，这是一种利用视频生成进行多模态推理的新范式。我们的VideoThinkBench显示，Sora-2在眼测谜题上比GPT5高出10%，在MMMU上准确率达到69%。

UMass-Embodied-AGI /Mirage

[CVPR 2026] Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens

UCSB-AI /GRIT

Official code for NeurIPS 2025 paper "GRIT: Teaching MLLMs to Think with Images"

EvolvingLMMs-Lab /OpenMMReasoner

[CVPR 2026] OpenMMReasoner: Pushing the Frontiers for Multimodal Reasoning with an Open and General Recipe

Cominclip /OmniVerifier

[ICLR 2026 Oral & ICML 2026 & NeurIPS 2026] Generative Universal Verifier as Multimodal Meta-Reasoner

The-Martyr /Awesome-Multimodal-Reasoning

Latest Advances on (RL based) Multimodal Reasoning and Generation in Multimodal LLMs

dlgjr /FINAR-VL

Open-source financial multimodal LLM for document understanding, chart/table reasoning, numerical reasoning, and financial analysis, with SFT, dual-branch RL, MOPD distillation, and public benchmark evaluation.

thuml /Reasoning-Visual-World

Official repository for "Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models", https://arxiv.org/abs/2601.19834

UCSB-AI /DMLR

[CVPR2026] Official codebase for the paper "Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space"

Wild-Cooperation-Hub /Awesome-MLLM-Reasoning-Benchmarks

A Comprehensive Survey on Evaluating Reasoning Capabilities in Multimodal Large Language Models.

Hai-chao-Zhang /ThinkJEPA

[NeurIPS 2026] ThinkJEPA: Empowering Latent World Models with Large Vision-Language Reasoning Model

rui-qian /READ

Rui Qian, Xin Yin, Dejing Dou†: Reasoning to Attend: Try to Understand How <SEG> Token Works (CVPR 2025)

AI4Phys /SeePhys

[NeurIPS 2025] Official implementation for the paper "SeePhys: Does Seeing Help Thinking? -- Benchmarking Vision-Based Physics Reasoning"

AdaptVision /AdaptVision

[CVPR 2026] AdaptVision: Efficient Vision-Language Models via Adaptive Visual Acquisition

yangjie-cv /WeThink

WeThink: Toward General-purpose Vision-Language Reasoning via Reinforcement Learning

zsxm1998 /awesome-perception-aware-rlvr

Unified reproductions, benchmarking & evaluation for perception-aware RLVR and OPD in VLMs, plus multi-turn agentic RL with HTTP, OpenEnv and OpenReward environments. 19 methods · 30+ benchmarks · Official code of CGPO (ACM MM 2026 Oral).

SceneCOT /scenecot

[ICLR 2026] SceneCOT: Eliciting Grounded Chain-of-Thought Reasoning in 3D Scenes

PKQZPC /HoloGeo

HoloGeo: Evidence-driven VLM for robust image geo-localization and landmark-bias mitigation.

ai4colonoscopy /Colon-X

Project COLON-X（塑造智能COLONoscopic领域的neXt前沿）

shawn0728 /ARES

该项目在 GitHub 上没有填写描述