ModelBeat
Models/MiniMax-M2.7
All models
M
minimax/minimax-m2.7

MiniMax-M2.7 is MiniMax’s open-weights AI model, released Mar 18, 2026. It costs $0.60/M input and $2.40/M output tokens, and handles a 205K-token context window.

MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.

Context
205K
Input
$0.60 / 1M
Output
$2.40 / 1M
Cheapest via DeepInfra
$0.25 / $1 / 1M
Released
Mar 18, 2026
Modalities
texttext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where MiniMax-M2.7 lands among all models tracked on Model Beat.

48.0
Intelligence Index
Epoch AI
48th percentile of tracked models
55.0
Coding Index
Epoch AI
55th percentile of tracked models
46.0
Agentic Index
Epoch AI
46th percentile of tracked models

Reasoning

3 evals
GPQA Diamond87.4%

Graduate-level scientific reasoning

Humanity's Last Exam28.1%

Frontier of human expert knowledge

WeirdML37.0%

Novel ML problem-solving

Coding

2 evals
SciCode47.0%

Scientific research coding

WebDev Arena1401

Human-rated web development

Agentic & Tools

2 evals
Terminal-Bench45.1%

Command-line agentic tasks

τ²-bench84.8%

Tool-agent-user reliability

In the news

1

Compare MiniMax-M2.7 with

Frequently asked questions

What is MiniMax-M2.7?

MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement.

Who created MiniMax-M2.7?

MiniMax-M2.7 is an AI model developed by MiniMax, part of the MiniMax family, released Mar 18, 2026.

How much does MiniMax-M2.7 cost?

MiniMax-M2.7's list price from MiniMax is $0.60 per million input tokens and $2.40 per million output tokens. The cheapest credible third-party provider on OpenRouter (DeepInfra) serves it at $0.25/$1 per 1M.

What is MiniMax-M2.7's context window?

MiniMax-M2.7 supports a context window of up to 205K tokens.

How does MiniMax-M2.7 perform on benchmarks?

On standardized evaluations tracked by Epoch AI, MiniMax-M2.7 scores 47.0% on SciCode, 87.4% on GPQA Diamond, 28.1% on Humanity's Last Exam.

Is MiniMax-M2.7 open source?

MiniMax-M2.7 is released with open weights (Open weights (non-commercial)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Jul 30, 2026.