mlsanityData to 14 Oct 2026

The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence

arXiv 2605.26494 ↗First posted 26 May 2026cs.AI cs.CL cs.LGcited by 61

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

also in 1 more
Abstract (arXiv)

We introduce the MiniMax-M2 series, a family of Mixture-of-Experts language models built around the principle that mini activations can unleash maximum real-world intelligence. The flagship M2 contains 229.9B total parameters with only 9.8B activated per token. Designed end-to-end for agentic deployment, the M2 series rests on three components: (i) agent-driven data pipelines producing large-scale, verifiable trajectories across agentic coding and agentic cowork, each grounded in an executable workspace and an artifact-aligned reward; (ii) Forge, a scalable agent-native RL system that adapts to long-horizon agent trajectories, paired with windowed-FIFO scheduling, prefix-tree merging, inference optimization, and a clean training-inference-agent decoupling that supports both white-box and black-box agents; (iii) the latest M2.7 checkpoint takes an early step toward self-evolution -- autonomously debugging training runs and modifying its own scaffold. Across M2 through M2.7, this combination translates a mini-activation footprint into frontier-tier performance on agentic coding, deep search, office-task, and reasoning benchmarks.

Key facts. Cited by 61 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 31 of its 45 references that are in the map. First posted on arXiv 26 May 2026.

Builds on

31 of 45 references.

References · 31

  1. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  2. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  3. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  4. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  5. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  6. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  7. 2020Longformer: The Long-Document Transformercited by 3,413
  8. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  9. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  10. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  11. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,848
  12. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
19 more
  1. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  2. 2023LongBench: A Bilingual, Multitask Benchmark for Long Context Understandingcited by 1,417
  3. 2023GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpointscited by 1,346
  4. 2024RULER: What's the Real Context Size of Your Long-Context Language Models?cited by 1,346
  5. 2023GAIA: a benchmark for General AI Assistantscited by 1,128
  6. 2024DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Modelscited by 867
  7. 2025BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agentscited by 647
  8. 2025τ2-Bench: Evaluating Conversational Agents in a Dual-Control Environmentcited by 627
  9. 2026Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfacescited by 585
  10. 2024Better & Faster Large Language Models via Multi-token Predictioncited by 360
  11. 2024Auxiliary-Loss-Free Load Balancing Strategy for Mixture-of-Expertscited by 205
  12. 2025Multi-SWE-bench: A Multilingual Benchmark for Issue Resolvingcited by 171
  13. 2024HELMET: How to Evaluate Long-Context Language Models Effectively and Thoroughlycited by 125
  14. 2025WebExplorer: Explore and Evolve for Training Long-Horizon Web Agentscited by 92
  15. 2023A Benchmark for Learning to Translate a New Language from One Grammar Bookcited by 75
  16. 2024KOR-Bench: Benchmarking Language Models on Knowledge-Orthogonal Reasoning Taskscited by 70
  17. 2024Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Modelscited by 57
  18. 2026DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Usecited by 13
  19. 2026CVE-Factory: Scaling Expert-Level Agentic Tasks for Code Security Vulnerabilitycited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 61.

61 papers.

Sort:

By area:

  1. 2026RLCSD: Reinforcement Learning with Contrastive On-Policy Self-Distillationcited by 45
  2. 2026Loop the Loopies!cited by 8
  3. 2026Cross-Tokenizer On-Policy Distillation via Byte-Prefix Marginalizationcited by 8
  4. 2026Beyond Function Calling: Benchmarking Tool-Using Agents under Tool-Environment Unreliabilitycited by 7
  5. 2026ICAE-Bench: Evaluating Coding Agents as Interactive Project Builderscited by 6
  6. 2026SHE: Trajectory-driven Safety Harness Evolution for LLM Agentscited by 6
  7. 2026HiMPO: Hindsight-Informed Memory Policy Optimization for Less-Entangled Credit in Long-Horizon Agentscited by 4
  8. 2026Self-Authored Verification Is Unreliable in Heuristic Self-Improving Agentscited by 4
  9. 2026SKT: Skill-Use Training at Scale via Verified Synthetic Data Generationcited by 3
  10. 2026LoopMTP: A looped transformer guided by latent multi-token predictioncited by 3
  11. 2026Ask Now, Use Later: Benchmarking the Proactivity Gap in Long-Lived LLM Agentscited by 2
  12. 2026VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Modelscited by 2
38 more
  1. 2026ACE: Pluggable Adaptive Context Elasticizer across Agentscited by 2
  2. 2026Antares: Foundation Models for Agentic Vulnerability Localizationcited by 2
  3. 2026Knowledge-Verified Emergent Deception in LLM Agents Under Conflicting Incentivescited by 2
  4. 2026Astar: Learning to Propose Evolution Directions for Self-Evolving Industrial AI Systemscited by 2
  5. 2026RingMoClaw: An Experience-Inspired Multi-Agent Framework for Self-Evolving Research in Remote Sensingcited by 2
  6. 2026Vortex: Efficient and Programmable Sparse Attention Serving for AI Agentscited by 1
  7. 2026Interactor: Agentic RL oriented Iterative Creation for Ad Description Generation in Sponsored Searchcited by 1
  8. 2026ReasoningLens: Hierarchical Visualization and Diagnostic Auditing for Large Reasoning Modelscited by 1
  9. 2026ExpertPlex: A High-Goodput Disaggregated Serving System for MoE LLMs with Adaptive Persistent Kernelscited by 1
  10. 2026SyncPlan: Long-Horizon LLM Coordination with Explicit Synchronization and Adaptive Correctioncited by 1
  11. 2026Active-SWE: Benchmarking Coding Agents for Proactive Bug Fixing without Issue Reportscited by 1
  12. 2026SemaPLC: A Project-Grounded, Verification-Gated Agent Harness for PLC Code Generationcited by 1
  13. 2026QATFactory: A Versatile, Deployment-Aligned Framework for Quantization-aware Training and Distillation of LLMscited by 1
  14. 2026OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agentscited by 0
  15. 2026Are LLMs Ready to Assist Physicians? PhysAssistBench for Interactive Doctor-Patient-EHR Assistancecited by 0
  16. 2026Token-Operations-Oriented Inference Optimization Techniques for Large Modelscited by 0
  17. 2026Beyond a Single Judge: The Evidence-Grounded, Social-Weighted Persona Panel for Generative UI Evaluationcited by 0
  18. 2026Would You Walk to the Car Wash? Revealing the Salience Bias of Large Language Models in Commonsense Reasoningcited by 0
  19. 2026MMShopBench: A Real-Log Benchmark for Multimodal, Multi-Turn Shopping Agentscited by 0
  20. 2026Wnuan: Staged Post-Training for Question Answering over Proprietary Enterprise Knowledgecited by 0
  21. 2026BrainBench: Benchmarking Large Language Models for Comprehensive EEG Understandingcited by 0
  22. 2026Benchmarking and Enhancing LLMs for Rule-Intensive Review of National Standard Documentscited by 0
  23. 2026Counterfactual Benchmarking and Training for Factuality Consistency and Order-Robust Grounded Reasoning in LLMs over Heterogeneous Knowledgecited by 0
  24. 2026Persuasive and Compliant Tendencies Predict Group Decision-Making in Humans and Language Modelscited by 0
  25. 2026TsuGO: Probing Search Efficiency in LLM Reasoning via Go Life-and-Death Problemscited by 0
  26. 2026LongDocBench: Benchmarking TOC Hierarchy and Contextual Relationship Recovery in Long Documentscited by 0
  27. 2026ClawSentry: A Progressive Multi-Tier Security Monitor for Safeguarding Autonomous LLM Agentscited by 0
  28. 2026SingProbe Technical Reportcited by 0
  29. 2026FiMI Banking: A Sovereign Model for Indian Retail Bankingcited by 0
  30. 2026DAREBench: Deployment-Aware and Reliable Evaluation of Models as Agentscited by 0
  31. 2026Vulnerability Localization Benchmark: Measuring Agentic Security Analysis at Repository Scalecited by 0
  32. 2026DENSE: Distilling Agent Trajectories into Evidence-Grounded Shortcut Trees for Self-Refinementcited by 0
  33. 2026TicTacBench: Benchmarking Timing Closure Capabilities of Coding Agentscited by 0
  34. 2026You Only Need 2/3 of the Chosen Experts: An Empirical Study of Dynamic Expert Pruning in Fine-Grained MoE LLMscited by 0
  35. 2026RECLAIM: Can Agents Reproduce the Claims of Machine Learning Papers?cited by 0
  36. 2026IndustryLLM: Failure-Driven LLM Training for Industrial Procurementcited by 0
  37. 2026IVT-Guard: All-in-One Reasoning Model for AI-Generated Content Detectioncited by 0
  38. 2026WavePP: High-Throughput Pipeline Parallel LLM Prefill under Prefix Reusecited by 0

50 most cited of 61 All 61 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025TinyLLM: Evaluation and Optimization of Small Language Models for Agentic Tasks on Edge Devices
  2. 2025MAPRO: Recasting Multi-Agent Prompt Optimization as Maximum a Posteriori Inference
  3. 2026Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale
  4. 2025Hephaestus: Improving Fundamental Agent Capabilities of Large Language Models through Continual Pre-Training
  5. 2026Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence
  6. 2026Qwen3-Coder-Next Technical Report
  7. 2026Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Model
  8. 2025R-Zero: Self-Evolving Reasoning LLM from Zero Data