MoonshotAI: Kimi K2.5

moonshotai/kimi-k2.5

Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed...

Prompt $/1M

$0.570

Completion $/1M

$2.850

Context

262,144 tok

Best provider

83 tok/s · ModelRun

Specs

Model id: moonshotai/kimi-k2.5
Display name: MoonshotAI: Kimi K2.5
Modality: text+image->text
Tokenizer: Other
Instruct type: —
Context length: 262,144
Max completion: 262,144
Prompt $/1M: $0.570
Completion $/1M: $2.850
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
ModelRun	83	422 ms	615 ms
BaseTen	71	795 ms	1,443 ms
Venicefastest	64	1,397 ms	2,893 ms
Ambient	63	684 ms	825 ms
Together	59	599 ms	881 ms
Artificial Analysis	49	1,067 ms	—
Parasail	47	611 ms	785 ms
Phala	44	1,619 ms	1,845 ms
Cloudflare	43	799 ms	1,063 ms
SiliconFlow	40	1,194 ms	1,593 ms
StreamLake	37	1,324 ms	1,600 ms
Inceptron	36	411 ms	597 ms
Moonshot AI	33	1,430 ms	1,987 ms
Novita	30	2,090 ms	2,602 ms
AtlasCloud	28	1,763 ms	2,331 ms
DeepInfra	25	862 ms	980 ms
NextBit	23	649 ms	764 ms
Io Net	17	1,550 ms	2,056 ms
Fireworks	12	6,185 ms	21,453 ms
DigitalOcean	7	1,212 ms	1,629 ms
Chutes	7	28,404 ms	34,648 ms