modelgrep
M

MiniMax: MiniMax M2.7

minimax/minimax-m2.7

15th smartest of 178ReasoningToolsJSON
Use via OpenRouter ↗
Intelligence
49.6
15th of 178
Design Elo
1287
Website
Speed
261
11th fastest
Latency
562ms
first token
Input price
$0.250
108th cheapest
Context
205K
131K max out

How it compares

Smarter than92%
of all ranked models
Faster than96%
of all ranked models
Cheaper than64%
of all ranked models

Overview

MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement. Built to actively participate in its own evolution, M2.7 integrates advanced agentic capabilities through multi-agent...

Benchmarks

independent · via OpenRouter
Artificial Analysis90th percentile
Intelligence Index
49.6
Coding Index
41.9
Agentic Index
61.5
GPQA Diamond
87%
Humanity's Last Exam
28%
SciCode
47%
Tau²-Bench (agentic)
85%
Design Arena · Elo12,223 tournaments
Website
1287
codecategories
1283
Game Dev
1275
Data Viz
1274
3D
1273
UI Component
1265
svg
1195
asciiart
1177

Providers & pricing (10)

ProviderIn $/MOut $/MUptime
DeepInfrafp8$0.250$1.00100%
Novitafp8$0.270$1.0899.9%
Morph$0.279$1.2097.9%
Fireworks$0.300$1.2099.9%
Togetherfp4$0.300$1.20100%
Minimaxfp8$0.300$1.2099%
Mara$0.300$1.2091.7%
Minimaxfp8$0.600$2.4094.7%
SambaNova$0.600$2.4099.3%
SambaNova$1.20$4.8088.7%

Specifications

Context window205K
Max output131K
Knowledge cutoff
Input modalitiestext
Output modalitiestext
Prompt caching
Cache read price$0.050/M
ModeratedNo

MiniMax M2.7 FAQ

How much does MiniMax M2.7 cost?

MiniMax M2.7 costs $0.250 per million input tokens and $1.00 per million output tokens via OpenRouter, making it 108th cheapest of 298 paid models.

How smart is MiniMax M2.7?

MiniMax M2.7 scores 49.6 on the Artificial Analysis Intelligence Index, ranking 15th of 178 benchmarked models, with a GPQA Diamond score of 87%.

How fast is MiniMax M2.7?

MiniMax M2.7 generates around 261 tokens per second with 562ms time-to-first-token (p50), the 11th fastest tracked model.

What is MiniMax M2.7's context window?

MiniMax M2.7 supports a 205K-token context window and can output up to 131K tokens. It accepts text input.

Compare head-to-head