Kimi API pricing
Every Kimi model we track, per 1M tokens, from the official pricing page.
Kimi API prices run from $0.95 in and $4 out per 1M tokens (Kimi K2.7 Code) up to $4 in and $20 out (Kimi K3.1 Pro Highspeed). Its cheapest model is number 47 of 77 on our full list. Checked Oct 29, 2026 against the official pricing page.
Moonshot AI's Kimi models: 6 are priced here, from $0.95 in and $4 out per 1M tokens (Kimi K2.7 Code) to $4 in and $20 out (Kimi K3.1 Pro Highspeed). Each model name below opens its own page with workload costs, LMArena ranks and cheaper alternatives.
| Model | Input | Cached | Output | Context |
|---|---|---|---|---|
| Kimi K2.7 Codekimi-k2.7-codeClosed to new sign-ups from Oct 31, 2026. Moonshot plans to retire the K2 series by the end of December 2026. | $0.95 | $0.19 | $4 | 262K |
| Kimi K2.6kimi-k2.6Closed to new sign-ups from Oct 31, 2026. Moonshot plans to retire the K2 series by the end of December 2026. | $0.95 | $0.16 | $4 | 262K |
| Kimi K2.7 Code Highspeedkimi-k2.7-code-highspeedAbout 180 tokens per second. Closed to new sign-ups from Oct 31, 2026. Moonshot plans to retire the K2 series by the end of December 2026. | $1.90 | $0.38 | $8 | 262K |
| Kimi K3.1 Prokimi-k3.1-proFlat price across the 1M window. Cache writes cost $2 (5-minute) or $4 (1-hour). | $2 | $0.20 | $10 | 1M |
| Kimi K3kimi-k3Previous flagship. Cache writes cost $3 (5-minute) or $6 (1-hour). | $3 | $0.30 | $15 | 1M |
| Kimi K3.1 Pro Highspeedkimi-k3.1-pro-highspeedSame model as K3.1 Pro at about 2 to 3 times the speed. Cache writes cost $4 (5-minute) or $8 (1-hour). | $4 | $0.20 | $20 | 1M |
Which Kimi model should you pick?
- Cheapest: Kimi K2.7 Code at $0.95 in and $4 out, number 47 of 77 on our full price list.
- Cheapest still open to new users: Kimi K3.1 Pro at $2 in and $10 out.
- Best on LMArena: Kimi K3.1 Pro, #6 of 50 priced models on Code Arena, at $2 in and $10 out.
- Flagship: Kimi K3 at $3 in and $15 out.
- Biggest context: Kimi K3.1 Pro with 1M tokens (3 models share it).
- Fast tier: Kimi K2.7 Code Highspeed costs 2x the output price of Kimi K2.7 Code ($8 vs $4).
- Fast tier: Kimi K3.1 Pro Highspeed costs 2x the output price of Kimi K3.1 Pro ($20 vs $10).
- Check the notes first: Kimi K2.7 Code, Kimi K2.6, Kimi K2.7 Code Highspeed are flagged as legacy, closed to new sign-ups or retiring on the official page.
How it compares
Cheapest models from other providers, blended per 1M tokens: Qwen3.7 Flash (Qwen) $0.06 · Step 3.5 Flash (StepFun) $0.15 · Qwen3.5 Flash (Qwen) $0.17. See the full API price table, the best model for each task or the plan vs API calculator.
Looking for the lowest price per token? See the cheapest LLM API ranking.
Head-to-head comparisons
FAQ
How much does the Kimi API cost?
From $0.95 input and $4 output per 1M tokens for Kimi K2.7 Code, up to $4 and $20 for Kimi K3.1 Pro Highspeed.
What is the cheapest Kimi model?
Kimi K2.7 Code, with a blended price of $1.71 per 1M tokens (3 parts input to 1 part output).
Is a subscription cheaper than the Kimi API?
It depends on how much you use it. Try our plan vs API calculator with your own usage.
Which Kimi model is the best?
By LMArena rank among the models we price, Kimi K3.1 Pro ranks highest at #6 on Code Arena.
Does the Kimi API have cheaper cached input?
Yes, for 6 of 6 models: cached input is listed at $0.16 to $0.38 per 1M tokens.
How does Kimi compare with other models?
We have head-to-head pages for Kimi K3.1 Pro vs Claude Opus 5.5, Kimi K3.1 Pro vs DeepSeek V4 Pro, Kimi K3.1 Pro vs GPT-6.1 Sol and 3 more.
Other providers: OpenAI · Anthropic · Google · Meta · DeepSeek · Qwen · Mistral · xAI · GLM (Z.ai) · MiniMax · ByteDance Seed · Tencent Hunyuan · Baidu ERNIE · StepFun · Xiaomi MiMo