NOUMENOpen app
NOUMEN

MiniMax · Open weights

MiniMax M2.5

Added Feb 12, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

MiniMax M2.5 is a productivity-focused flagship model that builds on M2.1 with stronger coding and real-world office workflow performance (Word, Excel, PowerPoint), plus better tool-use planning and token efficiency.

  • Reasoning
  • Tools
  • Structured output
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.27
Output
$1.08
Cache read
$0.027

Reference prices as published by the model provider. API example

Specifications

Context window
204.8K
Max output
131.1K
Average output (7 days)
653 tokens
Longer than 58% of 465 models

Side by side

MiniMax M2.5 compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

MiniMax M2.5MiniMaxThis modelMiniMax M3 ThinkingMiniMaxRemove
ProviderMiniMaxMiniMax
Discovery tierOpen weightsOpen weights
AddedFeb 12, 2026Jun 1, 2026
Limits
Context window204.8K512KBest
Max output131.07KBest80K
Average output653 tokens1.9K tokens
Pricing, USD per 1M tokens
Input$0.27$0.23Best
Output$1.08$0.96Best
Cache read$0.027Best$0.05
Performance
Throughput110.4 tok/s218.4 tok/sBest
Time to first token1.8 sBest7.2 s
Request success (24h)100.0%Best99.9%
Benchmark indexes
Intelligence Index22.829.2Best
Coding Index—58.6
Capabilities
ReasoningYesYes
ToolsYesYes
VisionNoYes
Structured outputYesYes
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • MiniMax M2.5
  • MiniMax M3 Thinking

Intelligence Index

  • MiniMax M2.522.8
  • MiniMax M3 Thinking29.2

MiniMax M2.5: Better than 55% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 58.6

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 21.3%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 4.4%

  • Harvey LAB-AAv1.1

    Legal tasks passed in full, with no material hallucination

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 0.0%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 1,089 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 1,243 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 9.8%

  • AA-LCRv1.1

    Reasoning across long documents

    • MiniMax M2.5: 73.3%
    • MiniMax M3 Thinking: 83.0%Top

  • MLCR-AA

    Long-context reasoning over medical records

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 4.4%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • MiniMax M2.5: 20.5%
    • MiniMax M3 Thinking: 39.0%Top

  • GPQA Diamond

    Graduate-level science questions

    • MiniMax M2.5: 84.8%
    • MiniMax M3 Thinking: 92.9%Top

  • IFBench

    Following precise output instructions

    • MiniMax M2.5: 71.6%
    • MiniMax M3 Thinking: 82.9%Top

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 2.0%

  • SciCode

    Scientific computing in Python

    • MiniMax M2.5: —
    • MiniMax M3 Thinking: 47.1%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • MiniMax M2.5: 34.8%
    • MiniMax M3 Thinking: 42.4%Top

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • MiniMax M2.5: 95.3%Top
    • MiniMax M3 Thinking: 88.9%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • MiniMax M2.5: 73.3%
    • MiniMax M3 Thinking: 83.0%Top

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
110.4 tok/s
Faster than 65% of 251 models
Time to first token
1.8 s
Quicker than 80% of 251 models
Prompt cache hits
2.9%
About $0.26 per 1M input tokens with caching
Request success (24h)
100.0%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 16.6%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax/minimax-m2.5",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="minimax/minimax-m2.5",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

MiniMax

Other MiniMax models, starting with the same family.