mlsanityData to 14 Oct 2026

Thyme: Think Beyond Images

arXiv 2508.11630 ↗First posted 15 Aug 2025cs.CVcited by 167

On the map: LLM reasoning → Multimodal reasoning

Abstract (arXiv)

Following OpenAI's introduction of the ``thinking with images'' concept, recent efforts have explored stimulating the use of visual information in the reasoning process to enhance model performance in perception and reasoning tasks. However, to the best of our knowledge, no open-source work currently offers a feature set as rich as proprietary models (O3), which can perform diverse image manipulations and simultaneously enhance logical reasoning capabilities through code. In this paper, we make a preliminary attempt in this direction by introducing Thyme (Think Beyond Images), a novel paradigm for enabling MLLMs to transcend existing ``think with images'' approaches by autonomously generating and executing diverse image processing and computational operations via executable code. This approach not only facilitates a rich, on-the-fly set of image manipulations (e.g., cropping, rotation, contrast enhancement) but also allows for mathematical computations, all while maintaining high autonomy in deciding when and how to apply these operations. We activate this capability through a two-stage training strategy: an initial SFT on a curated dataset of 500K samples to teach code generation, followed by a RL phase to refine decision-making. For the RL stage, we manually collect and design high-resolution question-answer pairs to increase the learning difficulty, and we propose GRPO-ATS (Group Relative Policy Optimization with Adaptive Temperature Sampling), an algorithm that applies distinct temperatures to text and code generation to balance reasoning exploration with code execution precision. We conduct extensive experimental analysis and ablation studies. Comprehensive evaluations on nearly 20 benchmarks show that Thyme yields significant and consistent performance gains, particularly in challenging high-resolution perception and complex reasoning tasks.

Key facts. Cited by 167 papers in the map (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 55 of its 63 references that are in the map. First posted on arXiv 15 Aug 2025.

Builds on

55 of 63 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025Skywork-R1V3 Technical Report8 Jul
  2. 2025High-Resolution Visual Reasoning via Multi-Turn Grounding-Based Reinforcement Learning8 Jul
  3. 2025Thinking with Generated Images28 May
  4. 2025Adaptive Chain-of-Focus Reasoning via Dynamic Visual Search and Zooming for Efficient VLMs21 May
  5. 2025DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning20 May
  6. 2025VisuLogic: A Benchmark for Evaluating Visual Reasoning in Multi-modal Large Language Models21 Apr
  7. 2025SoTA with Less: MCTS-Guided Sample Selection for Data-Efficient Visual Reasoning Self-Improvement10 Apr
  8. 2025VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model10 Apr
  9. 2025LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL10 Mar
  10. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models9 Mar
  11. 2024LogicVista: Multimodal LLM Logical Reasoning Benchmark in Visual Contexts6 Jul
  12. 2024We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?1 Jul
4 more
  1. 2024Visual CoT: Advancing Multi-Modal Language Models with a Comprehensive Dataset and Benchmark for Chain-of-Thought Reasoning25 Mar
  2. 2024Measuring Multimodal Mathematical Reasoning with MATH-Vision Dataset22 Feb
  3. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts3 Oct
  4. 2022ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning19 Mar

Other references · 39

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2025Qwen2.5-VL Technical Reportcited by 5,587
  5. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,719
  6. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,565
  7. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  8. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,729
  9. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 914
  10. 2024Show-o: One Single Transformer to Unify Multimodal Understanding and Generationcited by 752
  11. 2024MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?cited by 586
  12. 2024BLINK: Multimodal Large Language Models Can See but Not Perceivecited by 550
27 more
  1. 2023V*: Guided Visual Search as a Core Mechanism in Multimodal LLMscited by 482
  2. 2023OCRBench: On the Hidden Mystery of OCR in Large Multimodal Modelscited by 476
  3. 2024VLMEvalKit: An Open-Source Toolkit for Evaluating Large Multi-Modality Modelscited by 455
  4. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 392
  5. 2025Seed1.5-VL Technical Reportcited by 373
  6. 2024MME-RealWorld: Could Your Multimodal LLM Challenge High-Resolution Real-World Scenarios that are Difficult for Humans?cited by 292
  7. 2025VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interactioncited by 197
  8. 2024VITA: Towards Open-Source Interactive Omni Multimodal LLMcited by 173
  9. 2025LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Tokencited by 166
  10. 2024Multimodal ArXiv: A Dataset for Improving Scientific Comprehension of Large Vision-Language Modelscited by 139
  11. 2025Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawingcited by 136
  12. 2024LLaVA-Critic: Learning to Evaluate Multimodal Modelscited by 134
  13. 2025Baichuan-Omni-1.5 Technical Reportcited by 111
  14. 2024ZoomEye: Enhancing Multimodal LLMs with Human-Like Zooming Capabilities through Tree-Based Image Explorationcited by 103
  15. 2025R1-Omni: Explainable Omni-Multimodal Emotion Recognition with Reinforcement Learningcited by 100
  16. 2025MM-RLHF: The Next Step Forward in Multimodal LLM Alignmentcited by 83
  17. 2024Beyond LLaVA-HD: Diving into High-Resolution Large Multimodal Modelscited by 76
  18. 2024MME-Survey: A Comprehensive Survey on Evaluation of Multimodal LLMscited by 72
  19. 2025R1-Reward: Training Multimodal Reward Model Through Stable Reinforcement Learningcited by 65
  20. 2024MM-Vet v2: A Challenging Benchmark to Evaluate Large Multimodal Models for Integrated Capabilitiescited by 57
  21. 2025Kwai Keye-VL Technical Reportcited by 47
  22. 2025Long-VITA: Scaling Large Multi-modal Models to 1 Million Tokens with Leading Short-Context Accuracycited by 39
  23. 2024Debiasing Multimodal Large Language Models via Penalization of Language Priorscited by 37
  24. 2022Vision-Language Intelligence: Tasks, Representation Learning, and Large Modelscited by 28
  25. 2025Scaling and Beyond: Advancing Spatial Reasoning in MLLMs Requires New Recipescited by 22
  26. 2025DAMA: Data- and Model-aware Alignment of Multi-modal LLMscited by 14
  27. 2025Aligning Multimodal LLM with Human Preference: A Surveycited by 13

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 40; 2026: 127.

167 papers.

Sort:

By area:

  1. 2025DeepEyesV2: Toward Agentic Multimodal Modelcited by 125
  2. 2025Monet: Reasoning in Latent Visual Space Beyond Images and Languagecited by 82
  3. 2026Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillationcited by 77
  4. 2025Kwai Keye-VL 1.5 Technical Reportcited by 73
  5. 2026Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligencecited by 54
  6. 2026Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perceptioncited by 50
  7. 2025Reinforced Visual Perception with Toolscited by 38
  8. 2025TIR-Bench: A Comprehensive Benchmark for Agentic Thinking-with-Images Reasoningcited by 36
  9. 2025RealUnify: Do Unified Models Truly Benefit from Unification? A Comprehensive Benchmarkcited by 33
  10. 2025SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learningcited by 33
  11. 2025Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoningcited by 31
  12. 2025JarvisEvo: Towards a Self-Evolving Photo Editing Agent with Synergistic Editor-Evaluator Optimizationcited by 31
38 more
  1. 2025Skywork-R1V4: Toward Agentic Multimodal Intelligence through Interleaved Thinking with Images and DeepResearchcited by 27
  2. 2025Video-STAR: Reinforcing Open-Vocabulary Action Recognition with Toolscited by 26
  3. 2025V-Thinker: Interactive Thinking with Imagescited by 26
  4. 2025CodeV: Code with Images for Faithful Visual Reasoning via Tool-Aware Policy Optimizationcited by 26
  5. 2025Skyra: AI-Generated Video Detection via Grounded Artifact Reasoningcited by 25
  6. 2025Sketch-in-Latents: Eliciting Unified Reasoning in MLLMscited by 25
  7. 2026GenAgent: Scaling Text-to-Image Generation via Agentic Multimodal Reasoningcited by 24
  8. 2025A Survey on Agentic Multimodal Large Language Modelscited by 23
  9. 2025RewardMap: Tackling Sparse Rewards in Fine-grained Visual Reasoning via Multi-Stage Reinforcement Learningcited by 21
  10. 2025AdaTooler-V: Adaptive Tool-Use for Images and Videoscited by 21
  11. 2025Perceptual-Evidence Anchored Reinforced Learning for Multimodal Reasoningcited by 20
  12. 2025DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Modelscited by 19
  13. 2025Toward Effective Tool-Integrated Reasoning via Self-Evolved Preference Learningcited by 17
  14. 2026LensWalk: Agentic Video Understanding by Planning How You See in Videoscited by 17
  15. 2025CodeDance: A Dynamic Tool-integrated MLLM for Executable Visual Reasoningcited by 16
  16. 2026PyVision-RL: Forging Open Agentic Vision Models via RLcited by 16
  17. 2025BaseReward: A Strong Baseline for Multimodal Reward Modelcited by 15
  18. 2025Thinking with Programming Vision: Towards a Unified View for Thinking with Imagescited by 15
  19. 2026LaViT: Aligning Latent Visual Thoughts for Multi-modal Reasoningcited by 15
  20. 2025DeepSketcher: Internalizing Visual Manipulation for Multimodal Reasoningcited by 14
  21. 2026Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Modelscited by 14
  22. 2026Imagination Helps Visual Reasoning, But Not Yet in Latent Spacecited by 12
  23. 2026Thinking in Uncertainty: Mitigating Hallucinations in MLRMs with Latent Entropy-Aware Decodingcited by 12
  24. 2026VTC-Bench: Evaluating Agentic Multimodal Models via Compositional Visual Tool Chainingcited by 12
  25. 2025GeoVista: Web-Augmented Agentic Visual Reasoning for Geolocalizationcited by 11
  26. 2026V-Zero: Answer-Label-Free On-Policy Distillation with Contrastive Evidence Gating for Fine-Grained Visual Reasoningcited by 11
  27. 2025Catching the Details: Self-Distilled RoI Predictors for Fine-Grained MLLM Perceptioncited by 9
  28. 2025Look Less, Reason More: Rollout-Guided Adaptive Pixel-Space Reasoningcited by 9
  29. 2025VACoT: Rethinking Visual Data Augmentation with VLMscited by 9
  30. 2025COOPER: A Unified Model for Cooperative Perception and Reasoning in Spatial Intelligencecited by 9
  31. 2026What Does Vision Tool-Use Reinforcement Learning Really Learn? Disentangling Tool-Induced and Intrinsic Effects for Crop-and-Zoomcited by 9
  32. 2026Bridging Perception and Reasoning: Token Reweighting for RLVR in Multimodal LLMscited by 9
  33. 2025Revisiting the Necessity of Lengthy Chain-of-Thought in Vision-centric Reasoning Generalizationcited by 8
  34. 2025VG-Refiner: Towards Tool-Refined Referring Grounded Reasoning via Agentic Reinforcement Learningcited by 8
  35. 2025Deep But Reliable: Advancing Multi-turn Reasoning for Thinking with Imagescited by 8
  36. 2026Seeing Is Believing? A Benchmark for Multimodal Large Language Models on Visual Illusions and Anomaliescited by 8
  37. 2026PatchCue: Enhancing Vision-Language Model Reasoning with Patch-Based Visual Cuescited by 8
  38. 2026ResAdapt: Adaptive Resolution for Efficient Multimodal Reasoningcited by 8

50 most cited of 167 All 167 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Thinking Without Images: Internalizing Visual Manipulation with On-Policy Self-Distillation
  2. 2026Thinking in Pictures: A Systematic Benchmark for Reasoning-driven Image Generation
  3. 2026Retrieved Images as Visual Thought: Training-Free Multimodal In-Context Learning for the Open-vs-Closed Gap
  4. 2025Visual Agentic Reinforcement Fine-Tuning
  5. 2026Unified Thinker: A General Reasoning Modular Core for Image Generation
  6. 2025Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers
  7. 2025Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens
  8. 2025From Illusion to Intention: Visual Rationale Learning for Vision-Language Reasoning