Skip to content
aiqbench

MiniMax · model passport

MiniMax: MiniMax M2.5

minimax-m2-5
  • Open weights
  • Modified MIT licence
  • 228.7B params
  • 204,800 context
  • Released 2026-02-12(HF repo created)
10/27axes covered
Overall · LMArena
1,391#162 of 396
Best percentile
85.5IFBench (Artificial Analysis)
Cohort standing
51.8median 56.6

What the data says

Benchmarked on 10 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 85.5 pct — IFBench (Artificial Analysis) (71.6%)

Weakest axes

  • 13.4 pct — Factuality Arena (1,385)

Uneven: 72.1 percentile points between its best and worst axis — 85.5 pct on IFBench (Artificial Analysis) versus 13.4 pct on Factuality Arena.

  • IFBench (Artificial Analysis)85.5 pct71.6%Verify →
  • Coding62.7 pct1,446Verify →
  • LiveBench62.2 pct60.1%Verify →
  • Hard prompts61.4 pct1,418Verify →
  • Instruction following60.1 pct1,381Verify →
  • Overall59.3 pct1,391Verify →
  • Terminal-Bench50 pct42.7%Verify →
  • ARC-AGI-234.8 pct4.9%Verify →
  • WebDev Arena34.5 pct1,386Verify →
  • Factuality Arena13.4 pct1,385Verify →
  • Vision ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

How to run it

Run it yourself

RepositoryMiniMaxAI/MiniMax-M2.5
Parameters228.7B
ArchitectureMiniMaxM2ForCausalLM
LicenceModified MIT

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "MiniMaxAI/MiniMax-M2.5"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Use it hosted

API model IDminimax/minimax-m2.5
Context length204,800 tokens
Max output32,768 tokens
Input price$0.27 per 1M tokens
Cached input price$0.03 per 1M tokens
Output price$0.95 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin-pPresence penaltyReasoningRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-kTop logprobsTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"minimax/minimax-m2.5","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • minimax-m2.5 stands at 51.8 (raw mean 52.4) — measured on 10 of 27 axes.

    standingtemplateVerify →

  • minimax-m2.5 stands at 51.8 versus a cohort median of 56.6 across 106 models released within 180 days — below the peer median.

    cohorttemplateVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →