ZMIME
MiniMax · Released Feb 12, 2026

MiniMax-M2.5

Prior MiniMax coding model for agent workflows, office edits, and automation.

  • Open weights
  • Reasoning
  • Tool calling
Capability
146.7ECI · #66 of 148
Input
$0.30per 1M tokens
Output
$1.20per 1M tokens
Context
205K131K max output
Pricing

What it costs

Input
$0.30per million tokens
Output
$1.20per million tokens
Cached input
$0.03per million tokens
Cache write
$0.375per million tokens
Blended (3:1)
$0.525cheaper than 59% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$1.20
Team assistant25M input + 5M output tokens$13.50
Production app250M input + 50M output tokens$135.00

Official MiniMax (minimax.io) API price, as listed on models.dev.

Capability

Independent benchmarks

146.7Epoch Capabilities Index
#66 of 148 scored models
88Median 146167

The shaded band is Epoch AI’s confidence range (142.3–147.9); the tick marks the median scored model.

Scores from Epoch AI, run independently of MiniMax.

Specifications

The details

Lab
MiniMax
Released
Feb 12, 2026
Context window
204,800 tokens
Max output
131,072 tokens
Inputs
Text
Output
Text
Reasoning
Always on
Tool calling
Yes
Structured output
No
API model ID
MiniMax-M2.5on MiniMax (minimax.io)
Availability
21 API providerslisted on models.dev

MiniMax (minimax.io) model documentation

Lab-reported

What MiniMax claims

Published by the lab at launch. Settings vary, so compare these only with care.

BenchmarkScoreSettingSource
SWE-Bench Verified75.8 resolved—Source
SWE-Atlas Codebase QnA10.3—Source
SWE-Atlas Refactoring19.52—Source
SWE-Atlas Test Writing18.6—Source
Same lab

More from MiniMax

Questions

About MiniMax-M2.5

How much does MiniMax-M2.5 cost?

MiniMax-M2.5 costs $0.30 input / $1.20 output per million tokens (official MiniMax (minimax.io) API price). Cached input is $0.03 per million tokens, and writing to the cache costs $0.375. At a 3:1 input-to-output mix that is $0.525 per million tokens, cheaper than 59% of the 360 priced models we track.

What is the context window of MiniMax-M2.5?

MiniMax-M2.5 accepts up to 204,800 tokens per request and can write up to 131,072 tokens in one response.

How good is MiniMax-M2.5?

Epoch AI gives MiniMax-M2.5 a Capabilities Index score of 146.7 (likely range 142.3–147.9), ranking it #66 of 148 models Epoch has scored.

Is MiniMax-M2.5 open source?

Yes. MiniMax publishes the weights, so it can be downloaded and self-hosted.

What inputs does MiniMax-M2.5 support?

MiniMax-M2.5 accepts text and replies in text. It is a reasoning model, supports tool calling.

When was MiniMax-M2.5 released?

MiniMax released MiniMax-M2.5 on Feb 12, 2026.

What are the best alternatives to MiniMax-M2.5?

The closest current models from other labs on capability, price and release date are GPT-5.4 nano (OpenAI, ECI 145.8, $0.20 / $1.25), Qwen3.5 Plus (Alibaba (Qwen), ECI 146.8, $0.40 / $2.40), DeepSeek V3.2 (DeepSeek, ECI 146.3, $0.296 / $0.48) and Nemotron 3 Ultra 550B A55B (NVIDIA, ECI 146.2, $0.50 / $2.50).