Kimi K3.1 Pro is Moonshot AI's newest and most capable model: an open-weight, 2.8-trillion-parameter model that reads up to a million tokens at once and understands text, images and video. This guide explains what it can do, how it compares with Kimi K3, every way to use it, what it costs and where it shines.
104B active per token across 896 experts.
1,048,576 tokens with efficient sparse attention.
API input and output. Cached input $0.20.
Free to download from Hugging Face.
Specifications and prices from Moonshot AI's launch materials and API documentation.
K3.1 Pro builds on Kimi K3, released in July 2026, with the same size and context but better efficiency, accuracy and new 3D skills, at a lower price.
| Kimi K3 | Kimi K3.1 Pro | |
|---|---|---|
| API input / output | $3 / $15 | $2 / $10 |
| Cached input | $0.30 | $0.20 |
| Long-context attention | Full attention | Sparse attention in 24 layers |
| False statements (finance tests) | 23.4% | 8.3% |
| 3D, CAD and game engines | Basic | Blender, CAD, Unity, Unreal, Godot, 3D Make |
| Open weights | Yes (July 2026) | Yes (4 November 2026) |
| Highspeed version | — | $4 / $20, 2–3× faster |
The 1M-token context, thinking effort, multimodal input and the drop in false statements.
Holds a whole codebase or stack of long reports at once. Sparse attention keeps long inputs efficient.
Reads screenshots, charts, scanned pages, photos and video alongside text.
Low for quick answers, High for everyday work, Max for the hardest problems.
Plans multi-step tasks, browses, uses a computer and calls tools, scoring 79.6 on OSWorld 2.1.
65.7 on Terminal-Bench 4.0, close to the best closed models, and the engine behind Kimi Code's hardest jobs.
Writes Blender and CAD scripts, builds games in Unity, Unreal and Godot, and powers Kimi 3D Make.
Moonshot reports 14–20% gains on finance tests and far fewer false statements than K3.
Download, fine-tune and run it yourself under a permissive licence.
A third cheaper than K3 on the API, with cached input at $0.20 per million tokens.
Each word goes to a few of 896 experts, so only 104B of 2.8T parameters work at a time.
Terminal-Bench 4.0 score out of 100. Higher is better.
All scores are Moonshot's own, from the K3.1 Pro launch materials. Independent results may differ; treat small gaps as a tie.
| Where | How | Cost |
|---|---|---|
| Kimi app & kimi.com | Choose the newest model in chat | Uses monthly credits |
| Kimi Work | Desktop agent for Windows and Mac | Uses monthly credits |
| Kimi Code | Switch model for harder coding jobs | Kimi Code allowance |
| Kimi API | Model ID kimi-k3.1-pro | $2 / $10 per 1M |
| Hugging Face | Download open weights, self-host | Free (your hardware) |
| Cloud partners | For example Databricks | Partner pricing |
Chat with the newest model draws on credits; the free plan has a small allowance. Plans and credits
Six ways to use K3.1 Pro, from the Kimi app to your own servers.
| K3.1 Pro | K3.1 Pro Highspeed | |
|---|---|---|
| API input / output | $2 / $10 per 1M | $4 / $20 per 1M |
| Speed | Standard | 2–3× faster replies (Moonshot's figure) |
| What gets faster | — | Model output only, not tool calls or scripts |
| In Kimi Code | Standard quota | About 2× quota, from Allegretto |
| Best for | Most work, batch jobs, cost control | Interactive coding where you wait on every reply |
Research, coding, finance and 3D: four jobs where K3.1 Pro stands out.
Read these [number] reports and compare their main findings in a table. Cite the page for each point and list where they disagree.
Read this repository, then refactor [module] for readability without changing behaviour. Run the tests and explain each change.
Using the attached annual reports, build a three-year revenue and margin model in a table. Flag every assumption and cite its source.
Design a 3D [object] about [size] cm tall for 3D printing, with walls at least 2 mm thick. Explain how to export it.
Turn the attached screenshot into a responsive HTML and CSS page. Match the layout and keep everything in one file.
Plan and carry out this task step by step: [task]. Show the plan first and wait for my OK before making changes.
Set thinking to Max for the hardest of these. More Kimi prompts
https://api.moonshot.ai/v1.kimi-k3.1-pro and choose reasoning_effort.Thinking counts as output tokens. Cached input costs $0.20 per million. Kimi API guide and cost calculator
This is a summary, not legal advice. Read the full licence before commercial use.