Moonshot: Kimi K3.1 Pro

Kimi K3.1 Pro is Moonshot AI's post-trained successor to Kimi K3, built with substantially more post-training compute for long-horizon coding and knowledge work. It always reasons before answering, reads images with text, calls tools and returns structured output across a 1M-token context window.

Compare models

kimi-k3.1-pro

AvailableMoonshotChat

Input / Output: $2.00 / $10.00
Context: 1M
Max output: 128K
Modalities: VisionChat
Capabilities: Tool usePrompt CacheReasoning

About Kimi K3.1 Pro

Kimi K3.1 Pro is Moonshot AI's post-trained version of Kimi K3, built with substantially more post-training compute. Moonshot presents it as its flagship for long-horizon coding and end-to-end knowledge work. Thinking is always on, and the reasoning_effort setting picks its depth: low, high, or max, with max as the default. It supports tool calling, JSON schema output, and image input.

Where it works well

Moonshot describes it as K3 post-trained with substantially more compute, aimed at long-horizon coding and end-to-end knowledge work.
Thinking is always on, and reasoning_effort accepts low, high, or max, so one model covers routine turns and hard multi-step problems.
Tool calling and structured output that follows a JSON schema let agents call functions and return results your code can parse.
Images can accompany the text, so screenshots, charts, and diagrams can be part of a coding or analysis task.
Prompt caching is supported, which helps long sessions that resend the same large context.

When to choose another model

Thinking cannot be switched off and the default effort is max, so routine requests need a lower reasoning_effort to stay quick.
Kimi K3.1 Pro Highspeed serves the same model at about two to three times the speed, so it fits live, latency-bound work better.
It answers in text only; creating images or speech needs a different model.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.
02
Send your first request
Copy the example for your language and run it against the endpoint.
Text to textResponses API
POST/v1/responses
API format:
curl https://api.tokenlab.sh/v1/responses \ -H "Content-Type: application/json" \ -H "Authorization: Bearer sk-xxx" \ -d '{ "model": "kimi-k3.1-pro", "input": "Hello!" }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default

per 1M tokens

Official price: Input $2.00 / Output $10.00 / Cache read $0.20 / Cache write $2.00 / Cache write (1h) $4.00 / Text Input $2.00 / Text Output $10.00
Verified price: —
Discount: —

Prompt cache pricing

Cache read

Official price: $0.20
Verified price: —
Discount: —

Cache write

Official price: $2.00
Verified price: —
Discount: —

Cache write (1h)

Official price: $4.00
Verified price: —
Discount: —

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default	Input $2.00 / Output $10.00 / Cache read $0.20 / Cache write $2.00 / Cache write (1h) $4.00 / Text Input $2.00 / Text Output $10.00	—	—
Prompt cache pricing
Cache read	$0.20	—	—
Cache write	$2.00	—	—
Cache write (1h)	$4.00	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Nov 5, 2026, 11:52 PM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Kimi K3.1 Pro in Console with a prompt ready to edit or send.

Help me try kimi-k3.1-pro with a short message at /v1/responses. Show the reply, latency, and cost.

Use cases

Best for

Reasoning
Vision

Long-horizon coding agents
Let an agent plan, edit, run tests, and iterate across a large repository over a long session, with thinking depth set for each task.
End-to-end knowledge work
Turn a brief and a set of source documents into a finished analysis or report, calling tools to fetch data and check figures along the way.
Screenshot-based debugging
Send a screenshot of a broken page or a chart with your question and get a diagnosis grounded in what the image shows.
Schema-bound extraction
Pull fields from messy documents into JSON that follows your schema, so downstream code can use the result without a cleanup step.

Prompt examples

Here is the failing CI log and the repository layout. Find the root cause, write the fix, and list the tests you would run to confirm it.

Read these three vendor contracts and return a JSON object that follows this schema, with each party, the renewal date, and the termination clause.

This screenshot shows the checkout page with a misaligned total. Explain which CSS could cause it and propose the smallest fix.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What is Kimi K3.1 Pro?

Moonshot AI's post-trained version of Kimi K3, built with substantially more post-training compute. Moonshot presents it as its flagship for long-horizon coding and end-to-end knowledge work, with always-on thinking, tool calling, structured output, and image input.

How does reasoning effort work on Kimi K3.1 Pro?

Thinking is always on. You set its depth with the reasoning_effort parameter, which takes low, high, or max, and Moonshot's default is max. Use low for routine requests and keep max for hard planning, debugging, or analysis.

How is Kimi K3.1 Pro different from Kimi K3?

K3.1 Pro starts from Kimi K3 and adds substantially more post-training compute, which Moonshot directs at long-horizon coding and end-to-end knowledge work. Kimi K3 is the earlier release it builds on, so choose K3.1 Pro for new work in those areas.

When should I use Kimi K3.1 Pro Highspeed instead?

Pick Highspeed when response time matters most, such as interactive agents or live coding sessions. Moonshot lists the same intelligence and capabilities at about two to three times the speed, for a higher per-token price.

Does Kimi K3.1 Pro accept images and return structured output?

Yes. It accepts images alongside text, supports tool calling, and can return structured output that follows a JSON schema. The reply itself is text, so generate images or speech with another model.

How much does Kimi K3.1 Pro cost?

On TokenLab, Kimi K3.1 Pro costs Input $2.00 / Output $10.00 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Kimi K3.1 Pro?

Kimi K3.1 Pro accepts up to 1,048,576 tokens of context and returns up to 131,072 tokens in one response.

Which endpoint should Kimi K3.1 Pro use?

Use https://api.tokenlab.sh/v1/responses for Kimi K3.1 Pro. The request example below shows the matching code shape.

Which operations does Kimi K3.1 Pro support?

Kimi K3.1 Pro supports Text to text. Select an operation above to see its endpoint and request example.