MiniMax M1
Jun 2025
MiniMax M1Compare inon
Benchmarks of MiniMax M115
BenchmarkByScoreRank
No benchmark matches
Rankingconfiguration unstated?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperMiMo v2.6 ProunstatedOpenwins 67%-36%$0.43 / $0.87Memory (4-bit): 274 GB → 615 GB+4
BetterGPT-6 Lunamaxwins 63%-76%$0.068Function calling only with none · Chat Completions+3
OpenGLM 5.2maxwins 69%+153%$1.47Memory (4-bit): 274 GB → 452 GBCheaperGLM 5.3 FlashunstatedOpenwins 66%-72%$0.15 / $0.50Defaults to max · Self-hosted+4CheaperMiMo v2.5 ProunstatedOpenwins 66%-36%$0.43 / $0.87Memory (4-bit): 274 GB → 614 GBModelScoreCostPer taskWhat changes
MiMo v2.6 ProunstatedOpenwins 67%-36%$0.43 / $0.87Memory (4-bit): 274 GB → 615 GB+4GLM 5.3 FlashunstatedOpenwins 66%-72%$0.15 / $0.50Defaults to max · Self-hosted+4MiMo v2.5 ProunstatedOpenwins 66%-36%$0.43 / $0.87Memory (4-bit): 274 GB → 614 GB
Qwen3.7 Plusunstatedwins 66%-34%$0.32 / $1.28MiMo v2 Prounstatedwins 65%-36%$0.43 / $0.87ModelScoreCostPer taskWhat changes
GPT-6 Lunamaxwins 63%-76%$0.068Function calling only with none · Chat Completions+3GPT-5.6 Lunaxhighwins 63%-47%$0.085GPT-5.6 Terraxhighwins 65%+429%$0.63Prompts with >272K input tokensGPT-6.1 Solmaxwins 69%+371%$0.72Remove unsupported parameters+5
Muse Spark 1.2xhighwins 67%+135%$0.97Only auto tool_choice is supported+1ModelScoreCostPer taskWhat changes
GLM 5.2maxOpenwins 69%+153%$1.47Memory (4-bit): 274 GB → 452 GBGLM 5.3maxOpenwins 68%+153%$2.01Memory (4-bit): 274 GB → 452 GB+4GLM 5unstatedOpenwins 68%+82%$1 / $3.20Context: 1M → 203K+1GLM 5.1unstatedOpenwins 68%+153%$1.40 / $4.40Context: 1M → 203K+1MiMo v2.6 ProunstatedOpenwins 67%-36%$0.43 / $0.87Memory (4-bit): 274 GB → 615 GB+4ModelScoreCostPer taskWhat changes
MiniMax M3unstatedOpenwins 62%—$0.30 / $1.20Context: 1M → 524K+1MiniMax M2.7unstatedOpenwins 56%—$0.30 / $1.20Context: 1M → 205KMiniMax M2.5unstatedOpenwins 55%—$0.30 / $1.20Context: 1M → 205KMiniMax M2unstatedOpenwins 45%—$0.30 / $1.20Context: 1M → 205KMiniMax M3.1 Flashthinkingn/a—$0.22Complete model response must be appendedMiniMax M2.1thinkingOpenn/a—$0.30 / $1.20Context: 1M → 205KMiniMax M1 40kOpenn/a——MiniMax M1 80kOpenn/a—$0.55 / $2.20All MiniMax models →Run it yourself
estimateFits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✗ 48 GB GPU✗ 80 GB GPU
LicenceApache-2.0free for commercial use
History
6 versionsBenchmarks
all categoriesChat & Knowledge
Arena Text · overall1,364#117
Arena Text · Chinese1,360.1#122
Arena Text · English1,363#123
Arena Text · French1,370.1#101
Arena Text · German1,349.8#99
Arena Text · instruction following1,311.8#126
Arena Text · Japanese1,217#103
Arena Text · Korean1,265.9#106
Arena Text · non-English1,319.4#122
Arena Text · Polish1,329.4#99
Arena Text · Russian1,329.6#122
Arena Text · Spanish1,353.1#113
MMLU-Pro · official boardlab81.1
Arena Text · overall · full cohort1,342.6#121
Writing & Language
Arena Text · creative writing1,319#117
Arena Text · creative writing · full cohort1,297.5#120
FAQ
How much does MiniMax M1 cost?›
$0.40 per 1M input tokens and $2.20 per 1M output tokens.
Is MiniMax M1 free to use?›
There is no free API; the cheapest paid price is $0.40 per 1M input tokens. The weights are open, so you can run it yourself.
How good is MiniMax M1, and what is it best at?›
#117 of 196 in Chat & Knowledge, #117 of 195 in Writing & Language.
Is MiniMax M1 better than GPT-5 Mini?›
On Arena Text · creative writing, MiniMax M1 scores 1,319 vs 1,323 — about the same.
How can I use MiniMax M1?›
Through the API of Minimax, Novita.
What is the context window of MiniMax M1?›
1M tokens, with up to 40K output tokens.
When was MiniMax M1 released? Is there a newer version?›
Released 2025-06-13. It is the current version in its line.
Is MiniMax M1 open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run MiniMax M1 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can MiniMax M1 read images?›
No, it accepts text only.
What is MiniMax M1?›
MiniMax M1 is an open-weights language model from MiniMax, released Jun 2025.
What are the best alternatives to MiniMax M1?›
MiMo v2.6 Pro (cheaper), GPT-6 Luna (better), GLM 5.2 (open weights).