Kimi K3

Moonshot AIOpen weights1.05M context

Kimi K3 is an open-weights model from Moonshot AI with a 1.05M-token context window. 12 providers list it, from $2.70 per million input tokens and $13.50 per million output tokens (SiliconFlow).

Maker
Moonshot AI
Licence
Open weights
Context window
1.05M
Providers listed
12
Lowest input, per 1M
$2.70
Lowest output, per 1M
$13.50
Moonshot AI's own API, per 1M
$3.00 in · $15.00 out
Effort settings
Low, High, Max

Providers serving Kimi K3

Cheapest first, by the blended rate of three input tokens to one output token. Prices in US dollars per million tokens, as of 26 Oct 2026.

ProviderInputOutputContextToolsJSON modeOpenAI-compatibleModel ID
SiliconFlow
$2.70$13.501.05MYesYesYesmoonshotai/Kimi-K3
DeepInfra
$2.85$14.251.05MYesYesNomoonshotai/Kimi-K3
Alibaba Cloud
$3.00$15.001.05MYesYesYeskimi-k3
Arcee AI
$3.00$15.001.05MYesYesYesmoonshotai/kimi-k3
Baseten
$3.00$15.001.05MYesYesYesmoonshotai/Kimi-K3
DigitalOcean
$3.00$15.001.05MYesYesYeskimi-k3
Fireworks
$3.00$15.001.05MYesYesYesaccounts/fireworks/models/kimi-k3
Kimi
$3.00$15.001.05MYesYesYeskimi-k3
Nebius
$3.00$15.001.02MYesYesYesmoonshotai/Kimi-K3
Novita
$3.00$15.001.05MYesYesYesmoonshotai/kimi-k3
Together AI
$3.00$15.001.05MYesYesNomoonshotai/Kimi-K3
Amazon Bedrock
$3.30$16.501MYesYes—kimi-k3

Benchmark scores

Published scores only, each linked to its source. Compare every model in Bench rankings

BenchmarkDefaultLowHighMax
GPQA Diamond
PhD-level scientific reasoning
93.5%v84.8%91.9%93.1%
MMMU Pro
Visual reasoning
81.6%v———
Humanity's Last Exam
Reasoning & knowledge
43.5%v———
APEX-Agents
Agentic office work
———47.8%
τ³-Banking
Agentic tool use (banking)
———37.1%

More from Moonshot AI