Kimi’s Open Platform model list and pricing pages now include `kimi-k3.1-pro` and `kimi-k3.1-pro-highspeed`. Product-side availability was covered earlier in Kimi Adds K3.1 Pro. Below are the API IDs, rates, and access rules. Related reading: Kimi K3.1 Has Shown Up in Kimi Code, but There's No Release Yet.

The platform describes K3.1 Pro as a 3T-class model with native vision and a 1M-token context window. Highspeed is the same intelligence at about 2–3× the speed. Official posts and the tech blog call it a 3T-class open-weight line, with open weights planned for 2026-11-04; weights are not listed on Hugging Face yet.

Same-platform CNY rates (per MTok):

kimi-k3.1-pro            input ¥13.00 / output ¥68.00 / cache hit ¥1.30 / cache write 5 min ¥13.00 / 1 h ¥26.00
kimi-k3.1-pro-highspeed  input ¥26.00 / output ¥136.00 / cache hit ¥1.30 / cache write 5 min ¥26.00 / 1 h ¥52.00
kimi-k3                  input ¥20.00 / output ¥100.00 / cache hit ¥2.00 / cache write 5 min ¥20.00 / 1 h ¥40.00

USD rates on the international pricing page (per MTok):

kimi-k3.1-pro            input $2.00 / output $10.00 / cache hit $0.20 / cache write 5 min $2.00 / 1 h $4.00
kimi-k3.1-pro-highspeed  input $4.00 / output $20.00 / cache hit $0.20 / cache write 5 min $4.00 / 1 h $8.00
kimi-k3                  input $3.00 / output $15.00 / cache hit $0.30 / cache write 5 min $3.00 / 1 h $6.00

Peer flagship API rates (input / output per MTok; 2026-10-24, Beijing time):

kimi-k3.1-pro     ¥13 / ¥68     CNY   Kimi Open Platform chat pricing (CN)
gpt-6-astra       $10 / $50     USD   OpenAI API pricing (Std ≤272K)
Claude Opus 5.5   $4 / $20      USD   Anthropic Claude pricing
gemini-4-argon    $2 / $10      USD   Gemini API pricing
deepseek-v4-pro   $1.32 / $3.96 USD   DeepSeek API pricing
qwen3.8-max       ¥12 / ¥36     CNY   Alibaba Cloud Model Studio billing (China North 2)

¥ and $ stay in separate columns, with no currency conversion. For `gemini-4-argon`, the $2 / $10 figures are the promo rate through 2026-11-30; the page lists $4 / $20 afterward. `deepseek-v4-pro` figures are peak cache-miss. Highspeed rates appear only in the same-platform lists above.

On the China quickstart, calls unlock after a minimum ¥10 top-up; the ¥15 new-user voucher cannot be used for K3.1 Pro. On the international quickstart, the model unlocks after a successful top-up (minimum $1); that page does not mention the voucher rule.

Thinking mode stays on. Set top-level `reasoning_effort` to `low`, `high`, or `max` (default `max`).

OpenRouter now lists `moonshotai/kimi-k3.1-pro` at about $2 / $10 per 1M tokens on the standard provider (cache read $0.20), with a Highspeed provider at about $4 / $20 and a 1,048,576-token context. That listing is a hosted API route, not open weights. Hugging Face still has no K3.1 weights entry.

We have not yet tested calling Kimi K3.1 Pro or K3.1 Pro Highspeed via the API on real tasks; if you have, you can reach the site owner through the contact link in the footer.

Sources

  • Kimi Open Platform model list (CN): https://platform.kimi.com/docs/models.md
  • Kimi Open Platform model list (EN): https://platform.kimi.ai/docs/models.md
  • Kimi Open Platform chat pricing (CN): https://platform.kimi.com/docs/pricing/chat
  • Kimi Open Platform chat pricing (EN): https://platform.kimi.ai/docs/pricing/chat
  • Kimi K3.1 Pro quickstart (CN): https://platform.kimi.com/docs/guide/kimi-k3-1-pro-quickstart.md
  • Kimi K3.1 Pro quickstart (EN): https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart.md
  • OpenAI API pricing: https://platform.openai.com/docs/pricing
  • Anthropic Claude pricing: https://docs.anthropic.com/en/docs/about-claude/pricing
  • Gemini API pricing: https://ai.google.dev/gemini-api/docs/pricing
  • DeepSeek API pricing: https://api-docs.deepseek.com/quick_start/pricing
  • Alibaba Cloud Model Studio billing: https://help.aliyun.com/zh/model-studio/billing-for-model-studio
  • OpenRouter model page (moonshotai/kimi-k3.1-pro): https://openrouter.ai/moonshotai/kimi-k3.1-pro