Models & pricing / kimi-k3.1-pro · Moonshot
Moonshot

kimi-k3.1-pro API

Kimi K3.1 Pro is Moonshot’s multimodal reasoning model focused on long coding tasks, 3D work and financial-source research. Use kimi-k3.1-pro through XiuAI, distinguish it from the separate Highspeed version, and verify tool-loop behavior and actual usage.[1][2]

kimi-k3.1-pro

Model overview

Published provider specifications. Available features depend on the current API route.

Input
text / image / video
Output
Text
Context
1M
API routes
OpenAI-compatible API

Complex code and source analysis

The vendor emphasizes coding, 3D creation and multi-step financial-source analysis. Evaluate cross-file code changes, reference-guided modeling scripts or numerical checks across documents. Creating assets requires your professional tools to execute the scripts; financial conclusions require source and calculation checks.[1]

Compare against K3 on your complete task: whether tool loops continue, output is usable, corrections are reduced and current cost is acceptable. Vendor evaluations and examples explain development focus; they do not guarantee an engineering outcome or investment return through XiuAI.[1]

Standard version, thinking and complete messages

Kimi lists kimi-k3.1-pro-highspeed separately. It is not an automatic service tier of this standard ID; its speed, price or quota descriptions should not be applied to kimi-k3.1-pro. Check the current XiuAI catalog for other available versions.[1][2]

Thinking is always on in K3.1 Pro, with top-level reasoning_effort controlling depth. Return complete assistant messages for tool and multi-turn requests, and read streamed reasoning_content separately from final content instead of saving only the displayed answer.[1]

Task costs and vendor account terms

Interpret usage with the current input, output and cache rates on this page. Source material, tool results and output across turns contribute to total cost. Kimi’s direct-service deposits, coupons and rate tiers are vendor account terms; XiuAI uses its own balance, key permissions and catalog prices.[1][2]

API access

Using K3.1 Pro through XiuAI

Call kimi-k3.1-pro with an authorized XiuAI key through the current OpenAI-compatible endpoint. Check fields and tool-result replay on a short representative request before adding material or visual input. XiuAI provides model access, not hosted Blender, spreadsheets or customer workflows.[1][2]

Send a request using your existing API key.

POST /v1/chat/completions
curl 'https://router-api.xiu.ai/v1/chat/completions' \
  -H "Authorization: Bearer $XIUAI_API_KEY" \
  -H 'Content-Type: application/json' \
  --data-raw '{
  "model": "kimi-k3.1-pro",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}'
Read the API documentation

Frequently asked questions

Does standard Pro automatically use Highspeed?

Do not assume so. The vendor uses a separate model name for Highspeed, while this request ID is kimi-k3.1-pro. Use the actual XiuAI catalog and rates rather than borrowing another ID’s speed description.[1][2]

Does the financial-research use case make output decision-ready?

No. It can assist with source retrieval, numerical checks and organizing reasoning, but sources, formulas and assumptions still need review. Results from a vendor evaluation are not a guarantee that each answer is correct.[1]

Why retain complete assistant messages in tool calls?

K3.1 Pro documentation requires replaying the returned message intact in multi-turn and tool workflows. Extracting only content can lose necessary state; handle reasoning increments separately from the final answer.[1]

Related models

kimi-k3

Sources and updates

Specifications reference provider documentation; prices and availability come from the current XiuAI catalog. Specifications checked: 2026-11-11

Content reviewed: 2026-11-16

[1] Kimi K3.1 Pro documentation[2] Current XiuAI model and pricing data