mlsanityData to 14 Oct 2026

AIR: Adaptive Interleaved Reasoning with Code in MLLMs

arXiv 2606.23678 ↗First posted 22 Jun 2026cs.CV cs.AIcited by 0

On the map: LLM reasoning → Multimodal reasoning

also in 1 more
Abstract (arXiv)

Following the paradigm shift initiated by OpenAI o3, interleaved reasoning with code to enhance multimodal large language models (MLLMs) has become a pivotal research frontier. The existing literature focuses primarily on tool-use within vision-perception tasks. However, such approaches typically rely on predefined heuristics for visual manipulation and are inherently incapable of addressing numerical computation problems due to their exclusive focus on visual operations. This paper empowers MLLMs with adaptive interleaved reasoning capabilities through extended reinforcement learning training on code-augmented complex numerical computation tasks. To this end, we propose a comprehensive three-component solution consisting of: a two-stage cold-start data construction pipeline, data filtering strategies for RL dataset curation, and an adaptive tool-invocation strategy leveraging a group-constrained reward function for interleaved reasoning trajectories. Extensive experiments demonstrate that after Reinforcement Learning training with the group-constrained reward function, performance improves by an average of 6.1 percentage points (pp) on evaluation benchmarks. Specifically, the accuracy for interleaved reasoning samples increases by 9.9 pp, and the overall success rate of tool-use exceeds 95%. Our data and code are available at: https://github.com/CongHan0808/AIR.git.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 34 of its 45 references that are in the map. First posted on arXiv 22 Jun 2026.

Builds on

34 of 45 references.

Closest ancestors · 13

Same direction, newest first

  1. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency25 Aug
  2. 2025Thyme: Think Beyond Images15 Aug
  3. 2025VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning17 Jul
  4. 2025Metis-RISE: RL Incentivizes and SFT Enhances Multimodal Reasoning Model Learning16 Jun
  5. 2025MiMo-VL Technical Report4 Jun
  6. 2025VTool-R1: VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool Use25 May
  7. 2025DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning20 May
  8. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning10 Apr
  9. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
  10. 2025Visual-RFT: Visual Reinforcement Fine-Tuning3 Mar
  11. 2024DynaMath: A Dynamic Visual Benchmark for Evaluating Mathematical Reasoning Robustness of Vision Language Models29 Oct
  12. 2024LogicVista: Multimodal LLM Logical Reasoning Benchmark in Visual Contexts6 Jul
1 more
  1. 2024We-Math: Does Your Large Multimodal Model Achieve Human-like Mathematical Reasoning?1 Jul

Other references · 21

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  3. 2023Visual Instruction Tuningcited by 7,869
  4. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  5. 2025Qwen2.5-VL Technical Reportcited by 5,587
  6. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  7. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  8. 2025Qwen3-VL Technical Reportcited by 2,793
  9. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  10. 2024LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Modelscited by 1,464
  11. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  12. 2025Group Sequence Policy Optimizationcited by 835
9 more
  1. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Modelscited by 433
  2. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 392
  3. 2025Reinforcement Learning with Verifiable Rewards Implicitly Incentivizes Correct Reasoning in Base LLMscited by 289
  4. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Surveycited by 205
  5. 2025Agentic Reinforced Policy Optimizationcited by 172
  6. 2025SimpleTIR: End-to-End Reinforcement Learning for Multi-Turn Tool-Integrated Reasoningcited by 146
  7. 2025DeepEyesV2: Toward Agentic Multimodal Modelcited by 125
  8. 2025Agentic Reasoning and Tool Integration for LLMs via Reinforcement Learningcited by 110
  9. 2025SuperRL: Reinforcement Learning with Supervision to Boost Language Model Reasoningcited by 18

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Reinforced MLLM: A Survey on RL-Based Reasoning in Multimodal Large Language Models
  2. 2025Training-Free Reasoning and Reflection in MLLMs
  3. 2025Observe-R1: Unlocking Reasoning Abilities of MLLMs with Dynamic Progressive Reinforcement Learning
  4. 2025Perception-R1: Advancing Multimodal Reasoning Capabilities of MLLMs via Visual Perception Reward
  5. 2026See Further, Think Deeper: Advancing VLM's Reasoning Ability with Low-level Visual Cues and Reflection
  6. 2025Simple o3: Towards Interleaved Vision-Language Reasoning
  7. 2025Interleaved Reasoning for Large Language Models via Reinforcement Learning
  8. 2025PeRL: Permutation-Enhanced Reinforcement Learning for Interleaved Vision-Language Reasoning