Model FatigueLLMs under $1, measured — and who holds your data
Models

kimi-k3

Kimi K3

Made by Moonshot AI. Sold by 5 providers, from $2.83 per million input tokens.

Maker
Moonshot AI
Released
2026-07-16
Context
1,113K tokens
Max output
1,049K tokens
Input
text, image, video
Output
text
Open weights
Yes

Where to buy it

ProviderInput $/MOutput $/MBlended*Holds your promptTraining
Alibaba Cloud$2.83$14.13$5.65runs it 🇨🇳 Chinano training
DeepInfra$2.85$14.25$5.70runs it 🇺🇸 United Statesno training
Amazon Bedrock GLOBAL$3.00$15.00$6.00runs it 🇺🇸 United Statesno training
Moonshot AI$3.00$15.00$6.00runs it 🇨🇳 Chinano training
nebius$3.00$15.00$6.00runs it 🇳🇱 Netherlandsno training
Amazon Bedrock US$3.30$16.50$6.60runs it 🇺🇸 United Statesno training

Blended is our calculation: (3 × input + output) ÷ 4, US dollars per million tokens — most calls read far more than they write. Tiers that are cheaper because they train on your data are not used for the “from” price.

Compare with every other model →