MiniMax: MiniMax M3

minimax/minimax-m3

MiniMax M3 is an open-weight model from MiniMax, ranking #34 of 235 on the TryAii Score. It's strongest on knowledge (#13 on GPQA). It's served by 13 independent hosts from $0.3/$1.2 per million input/output tokens — about a quarter of what models at its level usually charge.

Prompt $/1M

$0.300

Completion $/1M

$1.200

Context

1,048,576 tok

Best provider

103 tok/s · SambaNova

Specs

Model id: minimax/minimax-m3
Display name: MiniMax: MiniMax M3
Modality: text+image+video->text
Tokenizer: Other
Instruct type: —
Context length: 1,048,576
Max completion: 512,000
Prompt $/1M: $0.300
Completion $/1M: $1.200
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionyes
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
SambaNovafastest	103	1,517 ms	1,838 ms
Venice	90	1,211 ms	1,627 ms
Artificial Analysis	85	1,166 ms	—
ModelRun	79	555 ms	1,110 ms
Minimax	78	1,012 ms	1,449 ms
CoreWeave	76	668 ms	809 ms
StreamLake	67	1,487 ms	1,839 ms
Novita	55	1,871 ms	2,495 ms
Together	48	1,172 ms	1,529 ms
Parasail	47	1,043 ms	1,751 ms
AtlasCloud	45	1,998 ms	2,560 ms
GMICloud	33	2,236 ms	2,882 ms
SiliconFlow	31	2,989 ms	3,746 ms
Morph	21	3,383 ms	4,696 ms
DeepInfra	15	1,698 ms	2,778 ms
minimaxfastest	—	—	—