Skip to main content

MiniMax: MiniMax M3

MiniMax: MiniMax M3 is minimax's model. On the un-leaked meo benchmark it ranks #14 of 22 with 50.2% accuracy and an Effective Value (𝕍) of 0.0001 β€” earning a correct answer for about $0.0074 in 129s each.

as of 2026-06-08

meo rank
#14
of 22
Accuracy
50.2%
overall
𝕍 index (top=100)
0.22
raw 𝕍 0.0001 Β· N=10
$ / correct
$0.0074
129s each

Per-domain accuracy

Framework bias
100%
Theory of mind
100%
CSP (unsatisfiable)
94%
Tape machine
93%
Regex automaton
88%
Perceptual illusions
55%
Watson-Glaser
25%
Logic / math / CS
8%
Base-rate bias
7%
Long arithmetic
7%
State tracking
0%

Efficiency

Cost per correct
$0.0074
Seconds per correct
129
Tokens per correct
5,956
Correct answers
112 / 251
Total benchmark cost
$0.83
Avg tokens / item
2,658

Aggregated third-party benchmarks

Artificial Analysis (artificialanalysis.ai). Redistribution requires an AA commercial license.

Model details

Lab
minimax
Released
2026-05-31
Weights
β€”
Context
1.0M
Max output
512K
Knowledge cutoff
β€”
Price in
$0.30/M
Price out
$1.20/M
Modality
text+image+video->text

Sources: openrouter Β· artificial_analysis Β· meo

FAQ

How accurate is MiniMax: MiniMax M3?

MiniMax: MiniMax M3 scores 50.2% overall on the meo benchmark (rank #14 of 22), an un-leaked private holdout graded objective-first.

What is MiniMax: MiniMax M3's Effective Value?

Its Effective Value (𝕍) is 0.0001 at N=10, Ο‰=1, Ξ΄=1.5 β€” a fused metric of accuracy, speed, cost, and error-cascade. 𝕍 rankings shift with task depth.

How much does MiniMax: MiniMax M3 cost per correct answer?

About $0.0074 per correct answer (5,956 tokens, 129 seconds each) at benchmark-time pricing via OpenRouter.

Is MiniMax: MiniMax M3 open weights?

Open-weights status for MiniMax: MiniMax M3 is not recorded.