Back

MiniMax

MiniMax-M2

ReasoningOpen weightsModels.devHugging Face
Intelligence
18.6
Rank 263 of 744 measured models
Output speed
64.7 t/s
Rank 265 of 403 measured models
Price/1M
$0.525
Rank 283 of 659 measured models
Context window
205K
Rank 612 of 1199 measured models

MiniMax-M2 scores 18.6 on the Artificial Analysis Intelligence Index, ranking 263rd of 744 measured models. The closest models on this index are JT-35B-Flash (18.7), Claude 4.1 Opus (18.6), and KAT-Coder-Pro V1 (18.6). At $0.525 per million tokens (3:1 input/output blended price), it is cheaper than Claude 4.1 Opus ($30).

Artificial Analysis Indices
Intelligence18.6
Rank 263 of 744 measured models
Math78.3
Rank 72 of 270 measured models
Standard Benchmarks
MMLU Pro82.0%
Rank 76 of 349 measured models
GPQA Diamond77.7%
Rank 231 of 619 measured models
HLE13.7%
Rank 275 of 664 measured models
LiveCodeBench82.6%
Rank 21 of 343 measured models
AIME 202578.3%
Rank 72 of 270 measured models
IFBench72.3%
Rank 50 of 450 measured models
AA-LCR64.3%
Rank 261 of 575 measured models
TerminalBench Hard25.8%
Rank 142 of 432 measured models
Additional benchmarks7
Design Arena · 3d48.3%
Design Arena · Codecategories48.1%
Design Arena · Dataviz50.0%
Design Arena · Gamedev48.1%
Design Arena · Svg55.3%
Design Arena · Uicomponent49.2%
Design Arena · Website48.0%
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy