Kimi K2 Thinking
As of Oct 4, 2026, the cheapest Kimi K2 Thinking API price is $0.47 per 1M input tokens (Vercel AI Gateway) and $2 per 1M output tokens (Vercel AI Gateway). Kimi K2 Thinking is offered by 18 providers in the kimi-thinking family, open weights, reasoning enabled.
API pricing by provider
Metered per-token prices. Cheapest input and output are highlighted; † marks a context-tier surcharge. How prices are ranked.
| Provider | 1M input | 1M output | 1M cache read | 1M cache write | Context |
|---|---|---|---|---|---|
| Vercel AI Gatewayopen | Cheapest$0.47 | Cheapest$2 | $0.141 | — | 216k |
| Helicone | $0.48 | $2 | — | — | 256k |
| IO.NET | $0.55 | $2.25 | $0.275 | $1.1 | 33k |
| Alibaba (China)open | $0.574 | $2.294 | — | — | 262k |
| 302.AI | $0.575 | $2.3 | — | — | 262k |
| Amazon Bedrockopen | $0.6 | $2.5 | — | — | 262k |
| Eden AIopen | $0.6 | $2.5 | — | — | 256k |
| Hugging Faceopen | $0.6 | $2.5 | $0.15 | — | 262k |
| Charm Hyperopen | $0.6 | $2.5 | $0.3 | — | 262k |
| Kilo Gatewayopen | $0.6 | $2.5 | — | — | 262k |
| DevPass (LLM Gateway)open | $0.6 | $2.5 | $0.06 | — | 262k |
| LLM Gatewayopen | $0.6 | $2.5 | $0.06 | — | 262k |
| NanoGPTopen | $0.6 | $2.5 | $0.15 | — | 262k |
| NovitaAIopen | $0.6 | $2.5 | $0.15 | — | 262k |
| OpenRouteropen | $0.6 | $2.5 | — | — | 262k |
| Meganovaopen | $0.6 | $2.6 | — | — | 262k |
Kimi K2 Thinking pricing FAQ
What is the cheapest Kimi K2 Thinking API?
As of Oct 4, 2026, Vercel AI Gateway has the lowest Kimi K2 Thinking output price at $2 per 1M tokens, and Vercel AI Gateway has the lowest input price at $0.47 per 1M tokens. Prices are metered per-token rates in USD; free tiers and subscription plans are excluded.
How many providers offer Kimi K2 Thinking?
18 providers list Kimi K2 Thinking on Sovyron; 16 of them sell it at a metered per-token price.
Is Kimi K2 Thinking free?
No provider in the Sovyron catalog lists a free tier for Kimi K2 Thinking.
What is the context window of Kimi K2 Thinking?
Kimi K2 Thinking supports a 262k-token context window and up to 262k output tokens.