MiniMax-M3

Release Date: 5/31/2026

Vals Index

Accuracy (Vals Index)

58.94% ± 1.82

Latency (Vals Index)

1477.67s

Cost/Test (Vals Index)

$1.50

Context Window

1M

Max Output Tokens

512k

Input Modality

Hyperparameter settings
Default Provider : MiniMax

Temperature

1

Top P

0.95

Top K

Default

Max Output Tokens

512,000

Show rankings only among open weight models

Benchmarks

Accuracy

Rankings

0.0%

± 1.82
6/ 23

0.0%

± 1.65
6/ 18

0.0%

± 0.92
4/ 109

0.0%

± 2.10
19/ 61

0.0%

± 1.96
2/ 61

0.0%

± 0.91
11/ 77

0.0%

± 3.94
14/ 35

0.0%

± 3.44
9/ 58

0.0%

± 0.86
39/ 115

0.0%

± 5.44
16/ 55

0.0%

± 1.44
4/ 109

0.0%

± 1.05
35/ 114

0.0%

± 0.44
8/ 112

0.0%

± 0.36
39/ 108

0.0%

± 0.94
24/ 74

0.0%

± 1.94
17/ 52
Contact us
Or send us an email at contact@vals.ai
Proprietary Benchmarks (contact us to get access)
Academic Benchmarks

Read about our methodology.