跳轉到主要內容

MiniMax M1 40k

MiniMaxMiniMax開源權重MIT · 商用許可

描述

MiniMax-M1 is an open-source, large-scale reasoning model that uses a hybrid-attention architecture for efficient long-context processing. It supports up to a 1 million token context window and 80,000-token reasoning output, matching Gemini 2.5 Pro’s scale while being highly cost-effective. Its Lightning Attention mechanism reduces compute requirements to about 30% of DeepSeek R1’s, and a new reinforcement learning algorithm, CISPO, doubles convergence speed compared to other RL methods. Trained on 512 H800s over three weeks, M1 achieves near state-of-the-art results across software engineering, long-context, and tool-use benchmarks, outperforming most open models and rivaling top closed systems.

發布日期
2025-06-17
參數規模
456.0B
上下文長度
—
支援模態
—

能力雷達圖

60
general
60
coding
70
reasoning
60
science估算
60
agents
10
multimodal

缺少專門科學評測時,Science 由 LLM Stats 科學得分或推理能力估算。

排行榜排名

暫無排名資料

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Chat

TAU-bench Retail67.8%自報
Multi-Challenge44.7%自報

Factuality

SimpleQA17.9%自報

Language

MMLU-Pro80.6%自報

Long Context

OpenAI-MRCR: 2 needle 128k76.1%自報
LongBench v261.0%自報
OpenAI-MRCR: 2 needle 1M58.6%自報

Math

MATH-50096.0%自報
AIME 202483.3%自報
AIME 202574.6%自報

Reasoning

ZebraLogic80.1%自報
GPQANYU + Cohere + Anthropic (2023)69.2%自報
LiveCodeBench62.3%自報
TAU-bench Airline60.0%自報
SWE-Bench Verified55.6%自報
Humanity's Last Exam7.2%自報

AA 評測指數

(Artificial Analysis)

暫無 AA 評測資料

LLM Stats 分類評分

(LLM Stats (zeroeval))
Language
80
Legal
80
Finance
80
Healthcare
80
Long Context
70
Math
70
Physics
70
Biology
70
Chemistry
70
Chat
60
Reasoning
60
Structured Output
60
Frontend Development
60
General
60
Code
60
Communication
60
Tool Calling
60
Factuality
20
Vision
10

定價

暫無定價資料

速度

暫無速度資料

供應商價格排行

暫無提供商資料

外部連結