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

• URL: http://arxiv.org/abs/2605.26494v1
• Date: Tue, 26 May 2026 03:16:11 GMT
• Title: The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence
• Abstract summary: MiniMax-M2 is a family of Mixture-of-Experts language models.<n>The flagship M2 contains 229.9B total parameters with only 9.8B activated per token.
• Score: 161.9185072343856
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• MCP-Cosmos: World Model-Augmented Agents for Complex Task Execution in MCP Environments [2.344185522187755]
  We present MCP-Cosmos, a framework that infuses generative World Models (WM) into the MCP ecosystem to enable predictive task automation.<n>We demonstrate that a "Bring Your Own World Model" (BYOWM) strategy allows agents to simulate state transitions and plans refine in a latent space before execution.
  arXiv  Detail & Related papers  (2026-05-09T19:25:56Z)
• Agent^2 RL-Bench: Can LLM Agents Engineer Agentic RL Post-Training? [23.212983812736343]
  Agent2 RL-Bench is a benchmark for evaluating agentic RL post-training.<n>We find that supervised pipelines dominate agent-driven post-training under fixed budgets.
  arXiv  Detail & Related papers  (2026-04-12T09:35:27Z)
• MagicAgent: Towards Generalized Agent Planning [73.21129030631421]
  We present textbfMagicAgent, a series of foundation models specifically designed for generalized agent planning.<n>We introduce a lightweight and scalable synthetic data framework that generates high-quality trajectories across diverse planning tasks.<n>We show that MagicAgent-32B and MagicAgent-30B-A3B achieve superior performance across diverse open-source benchmarks.
  arXiv  Detail & Related papers  (2026-02-22T01:39:16Z)
• Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models [78.73992315826035]
  We introduce Youtu-LLM, a lightweight language model that harmonizes high computational efficiency with native agentic intelligence.<n>Youtu-LLM is pre-trained from scratch to systematically cultivate reasoning and planning capabilities.
  arXiv  Detail & Related papers  (2025-12-31T04:25:11Z)
• LFM2 Technical Report [87.58431408281973]
  We present LFM2, a family of Liquid Foundation Models designed for efficient on-device deployment and strong task capabilities.<n>The LFM2 family covers 350M-8.3B parameters, including dense models (350M, 700M, 1.2B, 2.6B) and a mixture-of-experts variant (8.3B total, 1.5B active)<n>We build multimodal and retrieval variants: LFM2-VL for vision-latency tasks, LFM2-Audio for speech, and LFM2-ColBERT for retrieval.
  arXiv  Detail & Related papers  (2025-11-28T17:56:35Z)
• MATRIX: Multimodal Agent Tuning for Robust Tool-Use Reasoning [65.200259961515]
  We develop a vision-centric agent tuning framework that automatically synthesizes multimodal trajectories.<n>We also introduce Pref-X, a set of 11K automatically generated preference pairs.<n>Across three benchmarks, Agent-X, GTA, and GAIA, MATRIX consistently surpasses both open- and closed-source VLMs.
  arXiv  Detail & Related papers  (2025-10-09T17:59:54Z)
• Foam-Agent 2.0: An End-to-End Composable Multi-Agent Framework for Automating CFD Simulation in OpenFOAM [3.8790078849619074]
  We introduce Foam-Agent, a multi-agent framework that automates the entire end-to-end OpenFOAM workflow from a single natural language prompt.<n>The framework uses Model Context Protocol (MCP) to expose its core functions as discrete, callable tools.<n> evaluated on a benchmark of 110 simulation tasks, Foam-Agent achieves an 88.2% success rate with Claude 3.5 Sonnet.
  arXiv  Detail & Related papers  (2025-09-17T23:44:18Z)
• MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention [90.68287470551311]
  MiniMax-M1 is an open-weight, large-scale hybrid-attention reasoning model.<n>It supports a context length of 1 million tokens, 8x the context size of DeepSeek R1.<n>MiniMax-M1 is trained using large-scale reinforcement learning.
  arXiv  Detail & Related papers  (2025-06-16T15:08:02Z)
• MCP-Zero: Active Tool Discovery for Autonomous LLM Agents [13.005899769943442]
  We introduce MCP-Zero, an active agent framework that restores tool discovery autonomy to LLMs themselves.<n>Instead of overwhelming models with all available tools, MCP-Zero enables agents to actively identify capability gaps, and request specific tools on-demand.<n>We construct MCP-tools, a comprehensive dataset of 308 MCP servers and 2,797 tools from the official Model-Context-Protocol repository.
  arXiv  Detail & Related papers  (2025-06-01T15:48:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.