mlsanityData to 14 Oct 2026

G1: Bootstrapping Perception and Reasoning Abilities of Vision-Language Model via Reinforcement Learning

arXiv 2505.13426 ↗First posted 19 May 2025cs.CVcited by 29

On the map: LLM reasoning → Multimodal reasoning

also in 2 more
Abstract (arXiv)

Vision-Language Models (VLMs) excel in many direct multimodal tasks but struggle to translate this prowess into effective decision-making within interactive, visually rich environments like games. This ``knowing-doing'' gap significantly limits their potential as autonomous agents, as leading VLMs often performing badly in simple games. To address this, we introduce VLM-Gym, a curated reinforcement learning (RL) environment featuring diverse visual games with unified interfaces and adjustable, compositional difficulty, specifically designed for scalable multi-game parallel training. Leveraging VLM-Gym, we train G0 models using pure RL-driven self-evolution, which demonstrate emergent perception and reasoning patterns. To further mitigate challenges arising from game diversity, we develop G1 models. G1 incorporates a perception-enhanced cold start prior to RL fine-tuning. Our resulting G1 models consistently surpass their teacher across all games and outperform leading proprietary models like Claude-3.7-Sonnet-Thinking. Systematic analysis reveals an intriguing finding: perception and reasoning abilities mutually bootstrap each other throughout the RL training process. Source code including VLM-Gym and RL training are released at https://github.com/chenllliang/G1 to foster future research in advancing VLMs as capable interactive agents.

Key facts. Cited by 29 papers in the map (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 18 of its 44 references that are in the map. First posted on arXiv 19 May 2025.

Builds on

18 of 44 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025VLM Q-Learning: Aligning Vision-Language Models for Interactive Decision-Making6 May
  2. 2025VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model10 Apr
  3. 2025R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization13 Mar
  4. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
  5. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models9 Mar

Other references · 13

  1. 2025Qwen2.5-VL Technical Reportcited by 5,587
  2. 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 2,057
  3. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  4. 2023Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agentscited by 398
  5. 2024BALROG: Benchmarking Agentic LLM and VLM Reasoning On Gamescited by 146
  6. 2023SmartPlay: A Benchmark for LLMs as Intelligent Agentscited by 125
  7. 2024GTBench: Uncovering the Strategic Reasoning Limitations of LLMs via Game-Theoretic Evaluationscited by 120
  8. 2024Cradle: Empowering Foundation Agents Towards General Computer Controlcited by 117
  9. 2025FlowReasoner: Reinforcing Query-Level Meta-Agentscited by 63
  10. 2023Can Large Language Models Play Text Games Well? Current State-of-the-Art and Open Questionscited by 30
  11. 2025JARVIS-VLA: Post-Training Large-Scale Vision Language Models to Play Visual Games with Keyboards and Mousecited by 28
  12. 2024Can VLMs Play Action Role-Playing Games? Take Black Myth Wukong as a Study Casecited by 24
1 more
  1. 2025Are Large Vision Language Models Good Game Players?cited by 19

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 14; 2026: 15.

29 papers.

Sort:

By area:

  1. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Surveycited by 205
  2. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  3. 2025REASONING GYM: Reasoning Environments for Reinforcement Learning with Verifiable Rewardscited by 130
  4. 2026CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulationcited by 88
  5. 2025Spotlight on Token Perception for Multimodal Reinforcement Learningcited by 39
  6. 2025Vision-Zero: Scalable VLM Self-Improvement via Strategic Gamified Self-Playcited by 37
  7. 2025Explain Before You Answer: A Survey on Compositional Visual Reasoningcited by 28
  8. 2025Scaling Agentic Reinforcement Learning for Tool-Integrated Reasoning in VLMscited by 17
  9. 2025HoneyBee: Data Recipes for Vision-Language Reasonerscited by 13
  10. 2026VisGym: Diverse, Customizable, Scalable Environments for Multimodal Agentscited by 11
  11. 2025Reinforcement Learning for Large Model: A Surveycited by 9
  12. 2026Rethinking Token-Level Policy Optimization for Multimodal Chain-of-Thoughtcited by 9
17 more
  1. 2026PRPO: Perception-Reinforced Policy Optimization via Token-Level Dynamic Advantage Reshapingcited by 5
  2. 2026What does RL improve for Visual Reasoning? A Frankenstein-Style Analysiscited by 4
  3. 2025UAV-VL-R1: Generalizing Vision-Language Models via Supervised Fine-Tuning and Multi-Stage GRPO for UAV Visual Reasoningcited by 3
  4. 2026Discovering Process-Outcome Credit in Multi-Step LLM Reasoningcited by 3
  5. 2026Gym-V: A Unified Vision Environment System for Agentic Vision Researchcited by 3
  6. 2026Odysseus: Scaling VLMs to 100+ Turn Decision-Making in Games via Reinforcement Learningcited by 3
  7. 2026Towards Generalist Game Players: An Investigation of Foundation Models in the Game Multiversecited by 3
  8. 2026CAST: Game Solvers as Turn-Level Teachers for LLM Agentscited by 3
  9. 2025Thinking by Doing: Building Efficient World Model Reasoning in LLMs via Multi-turn Interactioncited by 2
  10. 2026RSICCLLM: A Multimodal Large Language Model for Remote Sensing Image Change Captioningcited by 2
  11. 2025More Than the Final Answer: Improving Visual Extraction and Logical Consistency in Vision-Language Modelscited by 1
  12. 2025Activating Visual Context and Commonsense Reasoning through Masked Prediction in VLMscited by 0
  13. 2025Training Multimodal Large Reasoning Models Needs Better Thoughts: A Three-Stage Framework for Long Chain-of-Thought Synthesis and Selectioncited by 0
  14. 2026Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Modelscited by 0
  15. 2026Chain-of-Adaptation: Surgical Vision-Language Adaptation with Reinforcement Learningcited by 0
  16. 2026Before Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoningcited by 0
  17. 2026ChartProbe: A Diagnostic Study on Visual Reasoning through Perception, Grounding, and Simple Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Game-RL: Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning
  2. 2025Perception Before Reasoning: Two-Stage Reinforcement Learning for Visual Reasoning in Vision-Language Models
  3. 2023Vision-Language Models as a Source of Rewards
  4. 2024Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning
  5. 2026VLGOR: Visual-Language Knowledge Guided Offline Reinforcement Learning for Generalizable Agents
  6. 2025Rethinking RL Scaling for Vision Language Models: A Transparent, From-Scratch Framework and Comprehensive Evaluation Scheme
  7. 2026All Roads Lead to Rome: Incentivizing Divergent Thinking in Vision-Language Models
  8. 2025VLA-R1: Enhancing Reasoning in Vision-Language-Action Models