whatmodel.sh / MiniMax-M1 (80k)

MiniMax-M1 (80k)

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

MiniMax-M1 (80k) is an open-weights language model from MiniMax, released June 2025. Its published size is 456B parameters (45.9B active per token) and a 1M-token context window. Licence: apache-2.0.

62.4±1.3#171 of 719 LLMsfrom 13 benchmark results across 2 sources

$ how it compares

The LLMs ranked either side of MiniMax-M1 (80k), out of 828 with a score, then more from MiniMax.

#modelscorewhereRAMreleased
199Trinity-Large-Thinking Arcee AI63±2api $0.39/M +local~141GB2026-04
200Magistral Medium 1.2 (2509) Mistral63±2comparison only2025-09
201MiniMax-M1 (80k) MiniMax62±1api $0.96/M +local~959GB2025-06
202Gemma 4 12B Google62±1local~8.5GB2026-06
203GPT-4.1 OpenAI62±1api $3.50/M2025-04
more from MiniMax
84MiniMax-M3 MiniMax70±1api $0.48/M +local~137GB2026-06
120MiniMax-M2.7 MiniMax68±1api $0.53/M +local~75GB2026-03
150MiniMax-M2.5 MiniMax66±1api $0.53/M +local~75GB2026-02

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks······█····· #1 of 1 * · 8.7% of its score · 3 tests
testresultvs the others in its rundifficultycountswhy
SWE-bench Verified lab56.0% ✎▁▁▁▁▁▁▂▄█▆▁▁ beats 45/131hard · avg 24%▇▇▇ 8.7%sharp · at its level · small run
τ-bench Airline (GPT-4.1 user) lab62.0% ✎█··········█ beats 1/1—not counted yet: no other model in the same run to compare it with
τ-bench Retail (GPT-4.1 user) lab63.5% ✎█··········█ beats 0/1—not counted yet: no other model in the same run to compare it with
coding······█····· #1 of 1 * · 13% of its score · 2 tests
testresultvs the others in its rundifficultycountswhy
SWE-bench Verified lab56.0% ✎▁▁▁▁▁▁▂▄█▆▁▁ beats 45/131hard · avg 24%▇▇▇ 8.7%sharp · at its level · small run
LiveCodeBench 2024-08–2025-05 (version unstated) lab65.0% ✎▃···█··▅··▃▃ beats 3/7medium · avg 42%▇▇ 4.3%at its level · small run
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #143 of 393 * · 14% of its score · 1 test

Beats INTELLECT-3, Nemotron 3 Super 120B-A12B, MiniMax-M2 here, though they rank higher overall. Trails LFM2.5 1.2B Thinking, LFM2.5 8B-A1B, LFM2.5 2.6B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
LMArena (text)1363.9 ✎▁▁▁▂▃▃▃▆▆▇█▃ beats 207/379avg 1268▇▇▇▇▇▇ 14%blunt
math······█····· #1 of 1 · 26% of its score · 3 tests
testresultvs the others in its rundifficultycountswhy
AIME 2025 lab76.9% ✎▁·▁▁▁▂▅▆█▇▅▁ beats 77/133medium · avg 42%▇▇▇▇ 10%sharp · small run
AIME 2024 lab86.0% ✎▁▁▁▂▁▃▁▅███▃ beats 61/85medium · avg 55%▇▇▇▇ 10%sharp · small run
MATH-500 lab96.8% ✎▁▃▂▃▅▇▅▇█▅▂▁ beats 64/84very easy · avg 89%▇▇ 6.1%maxed out · small run
other tests7 not in a task · 46% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond lab70.0% ✎▁▁▁▁▂▄▅▅███▄ beats 153/310hard · avg 43%▇▇▇▇▇ 12%sharp · at its level · small run
LongBench v2 lab61.5% ✎▂·▁·▂▂▁▃▄▁█▄ beats 24/28hard · avg 39%▇▇▇ 6.6%at its level · small run
Humanity's Last Exam lab8.4% ✎▁▁▁▃▆▆▇██▆▁▃ beats 11/104very hard · avg 3.4%▇▇▇ 6.5%small run
SimpleQA lab18.5% ✎▄▄▄▆█▆███·▂█ beats 19/32very hard · avg 7.7%▇▇ 5.8%small run
MMLU-Pro lab81.1% ✎▁▁·▁▁▁▃▂▄▅█▁ beats 131/188medium · avg 59%▇▇ 5.7%small run
ZebraLogic lab86.8% ✎█·▁▁▁▇·▄▄▁▄▁ beats 14/24medium · avg 60%▇▇ 5.6%small run
MultiChallenge lab44.7% ✎▁·▁·▁▃▆▅▃▁█▅ beats 9/29hard · avg 27%▇▇ 3.8%at its level · small run

$ where to run it

Through an API

api · Novita AI · $0.96/M · $0.55 in / $2.2 out per 1M tokens · 1M context · novita.ai ↗ · price sourceUSD per 1M text tokens, standard uncached input and output. Model endpoint minimaxai/minimax-m1-80k. Product: Serverless chat completions. API model: minimaxai/minimax-m1-80k. Settings: service tier=standard, batch=no, cache=uncached input. Additional/excluded charges: Tools, cache storage, priority service and other modalities excluded.. Currency: USD.

Other ways to run it

  1. Run it as published · needs ~912.21 GB · from huggingface.co/MiniMaxAI (official)
    pip install transformers torch accelerate && python -c "from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig; p='MiniMaxAI/MiniMax-M1-80k'; m=AutoModelForCausalLM.from_pretrained(p, device_map='auto', trust_remote_code=True); t=AutoTokenizer.from_pretrained(p, trust_remote_code=True); x=t(t.apply_chat_template([{'role': 'user', 'content': [{'type': 'text', 'text': 'Do you have mayonnaise recipes?'}]}], tokenize=False, add_generation_prompt=True), return_tensors='pt').to(m.device); o=m.generate(**x, generation_config=GenerationConfig(max_new_tokens=20, eos_token_id=t.eos_token_id, use_cache=True)); print(t.batch_decode([o[0][x.input_ids.shape[1]:]], skip_special_tokens=True)[0])"

Official weights and documentation: huggingface.co/MiniMaxAI ↗

$ how the score was worked out · MiniMaxAI/MiniMax-M1-80k mode… 85%, LMArena leaderboard dataset 14%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
MiniMaxAI/MiniMax-M1-80k model card · lab-reported12▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 85%apache-2.0
LMArena leaderboard dataset1▇▇▇▇ 14%CC BY 4.0
prior · pull toward the middle▇ 1%
weighted average= 62.4

not counted (2)

MiniMaxAI/MiniMax-M1-80k model cardτ-bench Airline (GPT-4.1 user) [correct]62.0%fewer than three models have this test yet, or this source ran it for this model alone
MiniMaxAI/MiniMax-M1-80k model cardτ-bench Retail (GPT-4.1 user) [correct]63.5%fewer than three models have this test yet, or this source ran it for this model alone

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json