MiniMax: MiniMax M2.5

minimax/minimax-m2.5

MiniMax-M2.5 is a SOTA large language model designed for real-world productivity. Trained in a diverse range of complex real-world digital working environments, M2.5 builds upon the coding expertise of M2.1...

Prompt $/1M

$0.150

Completion $/1M

$0.900

Context

204,800 tok

Best provider

172 tok/s · Mara

Specs

Model id: minimax/minimax-m2.5
Display name: MiniMax: MiniMax M2.5
Modality: text->text
Tokenizer: Other
Instruct type: —
Context length: 204,800
Max completion: 196,608
Prompt $/1M: $0.150
Completion $/1M: $0.900
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionno
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
Marafastest	172	988 ms	1,247 ms
Parasail	98	375 ms	439 ms
Friendli	94	840 ms	949 ms
AkashML	93	410 ms	697 ms
WandB	83	418 ms	515 ms
Artificial Analysis	82	1,022 ms	—
StreamLake	81	649 ms	743 ms
Nebius	73	874 ms	962 ms
AtlasCloud	70	1,305 ms	1,621 ms
DeepInfra	67	1,002 ms	1,178 ms
Baidu	63	3,114 ms	3,522 ms
Venice	60	1,502 ms	2,177 ms
SambaNova	56	3,074 ms	4,400 ms
Together	50	1,057 ms	1,231 ms
DigitalOcean	49	665 ms	786 ms
Novita	44	1,636 ms	1,840 ms
SiliconFlow	43	1,987 ms	2,226 ms
Minimax	43	2,548 ms	3,090 ms
Inceptron	40	738 ms	1,031 ms
Chutes	29	1,712 ms	1,826 ms
Phala	27	1,783 ms	1,899 ms
NextBit	27	2,720 ms	3,406 ms
Ionstream	19	802 ms	1,037 ms
Io Net	14	867 ms	1,104 ms