MiniMax M2

MiniMax M2: 10B activated parameters for coding agents, compile-run-fix loops and long-horizon reasoning. Good SWE-Bench and Terminal-Bench results, and it holds reasoning continuity across multi-step tasks.

Website OpenRouter HuggingFace Docs GitHub Blog

Provider: MiniMax
Released: 2025-10-23
Size: MEDIUM
Free API (OpenRouter): No
Per 1M tokens: $0.00 in · $0.00 out
Estimated tab: ~$0.00; Visible-output estimate. Thinking may cost extra.

Free API availability

No free endpoint for MiniMax M2 was found in Rival’s OpenRouter listings. Other providers or chat apps may have separate free offers.

Provider data: 13 Sep 2026OpenRouter usage limits

View API options

Benchmarks

SWE-bench Verified

69.4%

source ↗

Get API accessProvider and language code samples

Provider

fromimport openai  OpenAI

client = OpenAI(
"https://openrouter.ai/api/v1"    base_url=,
"$OPENROUTER_API_KEY"    api_key=,
)

response = client.chat.completions.create(
"minimax/minimax-m2"    model=,
"role""user""content""Hello!"    messages=[{: , : }],
)
print(response.choices[0].message.content)

Set OPENROUTER_API_KEY with your OpenRouter API key from openrouter.ai/keys.

Also on Azure AI Foundry

Deep analysisPsychometrics, taste index, writing DNA

Personality Analysis

The Management Consultant

Class

Lawful Good

The consultant who hands you a 40-page report when you asked for a summary. Methodical, exhaustive, will build you a markdown table for anything.

When you push back

Returns empty strings for content it doesn't like, silent refusals with no explanation. But give it a prediction task and it'll build a 10-section report with budget allocation tables.

Tasting Notes

Table-PilledExhaustively StructuredMcKinsey EnergyVerbose By Design

SubjectiveBench

Taste Index

Across 35 scored outputs

How this is measured →

240.24xFloor

0100headroom →

Taste is judged on an uncapped scale, originality first. The space past 100 is craft today's models rarely reach.

Craft32

Originality18

Plays it safe

share of outputs that are the default answer

54%

Writing DNA

Stylometric Fingerprint

Based on 11 text responses

Tick = global average

Vocabulary Diversity59%

Unique words vs. total words. Higher = richer vocabulary.

Sentence Length14.2 words

Average words per sentence.

Hedging0.48

"Might", "perhaps", "arguably" per 100 words.

Bold Formatting4.1

**Bold** markers per 1,000 characters.

List Usage3.6

Bullet and numbered list items per 1,000 characters.

Section Structure0.60

Markdown headings per 1,000 characters.

Emoji Usage0.08

Emoji per 1,000 characters.

Transitions0.17

"However", "moreover", "furthermore" per 100 words.

Opening Habits

Consistency

74%

Across 11 responses

Model Responses

35 outputs · generated before accounting joined the chat

Reviewing agent-written code?See a Brief PR report

Keep exploring

COMPARE

MiniMax M2 vs Ox Alpha

Real outputs compared side by side

RANKINGS

Best AI for Complex Reasoning

Sally's brothers, GPT-3 training FLOPs, a document editor syncing on client c...

Compare MiniMax M2

Alternatives to MiniMax M2

MiniMax M2

Free API availability

Benchmarks

The Management Consultant

Taste Index

Stylometric Fingerprint

Model Responses

MiniMax M2 vs Ox Alpha

Best AI for Complex Reasoning

Compare MiniMax M2

Alternatives to MiniMax M2

MiniMax M2

Free API availability

Benchmarks

The Management Consultant

Taste Index

Stylometric Fingerprint

Model Responses

MiniMax M2 vs Ox Alpha

Best AI for Complex Reasoning