mlsanityData to 14 Oct 2026

Kimi K2: Open Agentic Intelligence

arXiv 2507.20534 ↗First posted 28 Jul 2025cs.LG cs.AI cs.CLcited by 567

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

also in 1 more
Abstract (arXiv)

We introduce Kimi K2, a Mixture-of-Experts (MoE) large language model with 32 billion activated parameters and 1 trillion total parameters. We propose the MuonClip optimizer, which improves upon Muon with a novel QK-clip technique to address training instability while enjoying the advanced token efficiency of Muon. Based on MuonClip, K2 was pre-trained on 15.5 trillion tokens with zero loss spike. During post-training, K2 undergoes a multi-stage post-training process, highlighted by a large-scale agentic data synthesis pipeline and a joint reinforcement learning (RL) stage, where the model improves its capabilities through interactions with real and synthetic environments. Kimi K2 achieves state-of-the-art performance among open-source non-thinking models, with strengths in agentic capabilities. Notably, K2 obtains 66.1 on Tau2-Bench, 76.5 on ACEBench (En), 65.8 on SWE-Bench Verified, and 47.3 on SWE-Bench Multilingual -- surpassing most open and closed-sourced baselines in non-thinking settings. It also exhibits strong capabilities in coding, mathematics, and reasoning tasks, with a score of 53.7 on LiveCodeBench v6, 49.5 on AIME 2025, 75.1 on GPQA-Diamond, and 27.1 on OJBench, all without extended thinking. These results position Kimi K2 as one of the most capable open-source large language models to date, particularly in software engineering and agentic tasks. We release our base and post-trained model checkpoints to facilitate future research and applications of agentic intelligence.

Key facts. Cited by 567 papers in the map (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 73 of its 91 references that are in the map. First posted on arXiv 28 Jul 2025.

Builds on

73 of 91 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains17 Jun

Other references · 72

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2022Training language models to follow instructions with human feedbackcited by 15,971
  4. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  5. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  6. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  7. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  8. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  9. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
  10. 2021Program Synthesis with Large Language Modelscited by 3,637
  11. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  12. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
38 more
  1. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  2. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  3. 2024LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Codecited by 2,004
  4. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
  5. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,947
  6. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,848
  7. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
  8. 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 1,631
  9. 2023Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generationcited by 1,391
  10. 2019ZeRO: Memory Optimizations Toward Training Trillion Parameter Modelscited by 1,301
  11. 2023Instruction-Following Evaluation for Large Language Modelscited by 1,228
  12. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  13. 2024DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Modelcited by 1,135
  14. 2023WizardLM: Empowering large pre-trained language models to follow complex instructionscited by 1,125
  15. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 940
  16. 2019DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphscited by 903
  17. 2024MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategiescited by 723
  18. 2023AGIEval: A Human-Centric Benchmark for Evaluating Foundation Modelscited by 716
  19. 2023C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Modelscited by 708
  20. 2023Scaling Vision Transformers to 22 Billion Parameterscited by 694
  21. 2022Large Language Models Can Self-Improvecited by 664
  22. 2021Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LMcited by 632
  23. 2025τ2-Bench: Evaluating Conversational Agents in a Dual-Control Environmentcited by 627
  24. 2023YaRN: Efficient Context Window Extension of Large Language Modelscited by 601
  25. 2024From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipelinecited by 500
  26. 2025Humanity's Last Examcited by 491
  27. 2023CMMLU: Measuring massive multitask language understanding in Chinesecited by 449
  28. 2025Muon is Scalable for LLM Trainingcited by 428
  29. 2024LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitaskscited by 379
  30. 2024Measuring short-form factuality in large language modelscited by 373
  31. 2025PaperBench: Evaluating AI's Ability to Replicate AI Researchcited by 321
  32. 2024CRUXEval: A Benchmark for Code Reasoning, Understanding and Executioncited by 300
  33. 2025SWE-smith: Scaling Data for Software Engineering Agentscited by 285
  34. 2025ZeroSearch: Incentivize the Search Capability of LLMs without Searchingcited by 208
  35. 2025ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoningcited by 201
  36. 2024Are We Done with MMLU?cited by 194
  37. 2023Small-scale proxies for large-scale Transformer training instabilitiescited by 183
  38. 2025MultiChallenge: A Realistic Multi-Turn Conversation Evaluation Benchmark Challenging to Frontier LLMscited by 180

50 most cited of 72 All 72 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 146; 2026: 421.

567 papers.

Sort:

By area:

  1. 2026DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligencecited by 1,040
  2. 2026Kimi K2.5: Visual Agentic Intelligencecited by 501
  3. 2026Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Modelscited by 455
  4. 2025Kimi Linear: An Expressive, Efficient Attention Architecturecited by 196
  5. 2026MiMo-V2-Flash Technical Reportcited by 191
  6. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  7. 2025ACON: Optimizing Context Compression for Long-horizon LLM Agentscited by 136
  8. 2025MCP-Bench: Benchmarking Tool-Using LLM Agents with Complex Real-World Tasks via MCP Serverscited by 109
  9. 2025MCP-Universe: Benchmarking Large Language Models with Real-World Model Context Protocol Serverscited by 96
  10. 2026SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalizationcited by 94
  11. 2025CWM: An Open-Weights LLM for Research on Code Generation with World Modelscited by 93
  12. 2025MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scalingcited by 81
38 more
  1. 2025A Survey on Diffusion Language Modelscited by 78
  2. 2026Qwen3-Coder-Next Technical Reportcited by 71
  3. 2025Baichuan-M2: Scaling Medical Capability with Large Verifier Systemcited by 59
  4. 2025SWE-EVO: Benchmarking Coding Agents in Long-Horizon Software Evolution Scenarioscited by 59
  5. 2025Simulating Environments with Reasoning Models for Agent Trainingcited by 57
  6. 2026Ultralytics YOLO26: Unified Real-Time End-to-End Vision Modelscited by 57
  7. 2025Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Modelcited by 55
  8. 2026Self-Distilled Agentic Reinforcement Learningcited by 55
  9. 2025RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environmentscited by 50
  10. 2026A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agentscited by 50
  11. 2025Diffusion Language Models are Super Data Learnerscited by 48
  12. 2025Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Modelscited by 48
  13. 2025Earth-Agent: Unlocking the Full Landscape of Earth Observation with Agentscited by 46
  14. 2025REAP the Experts: Why Pruning Prevails for One-Shot MoE compressioncited by 44
  15. 2025Budget-Aware Tool Use Enables Effective Agent Scalingcited by 44
  16. 2026Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoningcited by 42
  17. 2025Kimi-Dev: Agentless Training as Skill Prior for SWE-Agentscited by 41
  18. 2025LLM-as-a-Prophet: Understanding Predictive Intelligence with Prophet Arenacited by 40
  19. 2025Spatia: Video Generation with Updatable Spatial Memorycited by 40
  20. 2025Toward Training Superintelligent Software Agents through Self-Play SWE-RLcited by 40
  21. 2026Nemotron 3 Super: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoningcited by 38
  22. 2025A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontierscited by 37
  23. 2025MuonBP: Faster Muon via Block-Periodic Orthogonalizationcited by 35
  24. 2025Scaling Behavior of Discrete Diffusion Language Modelscited by 35
  25. 2026Reinforcement-aware Knowledge Distillation for LLM Reasoningcited by 35
  26. 2026On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stabilitycited by 35
  27. 2026ScaleEnv: Scaling Environment Synthesis from Scratch for Generalist Interactive Tool-Use Agent Trainingcited by 32
  28. 2026SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecturecited by 32
  29. 2025SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agentscited by 31
  30. 2025ChatInject: Abusing Chat Templates for Prompt Injection in LLM Agentscited by 31
  31. 2025ToolMind Technical Report: A Large-Scale, Reasoning-Enhanced Tool-Use Datasetcited by 30
  32. 2026SkillCraft: Can LLM Agents Learn to Use Tools Skillfully?cited by 30
  33. 2025BudgetThinker: Empowering Budget-aware LLM Reasoning with Control Tokenscited by 29
  34. 2025SPELL: Self-Play Reinforcement Learning for Evolving Long-Context Language Modelscited by 29
  35. 2026TermiGen: High-Fidelity Environment and Robust Trajectory Synthesis for Terminal Agentscited by 29
  36. 2025AI Agentic Programming: A Survey of Techniques, Challenges, and Opportunitiescited by 28
  37. 2026GameWorld: Towards Standardized and Verifiable Evaluation of Multimodal Game Agentscited by 28
  38. 2026Look Inward to Explore Outward: Learning Temperature Policy from LLM Internal States via Hierarchical RLcited by 27

50 most cited of 567 All 567 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Kimi K3: Open Frontier Intelligence
  2. 2026A.X K1 Technical Report
  3. 2026A.X K2 Technical Report
  4. 2025Klear-AgentForge: Forging Agentic Intelligence through Posttraining Scaling
  5. 2025K2-Think: A Parameter-Efficient Reasoning System
  6. 2026OpenThoughts-Agent: Data Recipes for Agentic Models
  7. 2025Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers
  8. 2025Making Qwen3 Think in Korean with Reinforcement Learning