None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
kimi k3.1
Try kimi k3.1 for code, 3D assets and sites — an openly released three-trillion-scale model with space for a million tokens and a fast-response build.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
What Kimi K3.1 Brings to Your Builds
Kimi K3.1 pairs an openly released three-trillion-scale model with space for a million tokens and stronger coding, 3D and finance performance.
- Three-trillion-scale open releaseThis K3.1 flagship is Kimi's largest openly released model to date, with a context window that holds a million tokens.
- Fast variant for quicker turnsA speed-focused companion returns answers 2-3x faster than the standard build, so back-and-forth work feels more immediate.
- Broad joint reinforcement learningTrained on K3 with much heavier post-training, it runs large-scale joint RL across hundreds of downstream areas, improving each one.
Getting Started with Kimi K3.1
Go from your first prompt to a finished build in three simple steps with Kimi K3.1.
What Kimi K3.1 Can Do
From long coding runs to parametric 3D, finance analysis and safety limits, these are the areas where Kimi K3.1 delivers.
Coding agents for long tasks
On Terminal-Bench 4.0 it scores 65.7%, ahead of GPT-6 Astra (58.2%) and close to Claude Opus 5.5 (66.4%); Kimi Code Bench 3.0 Long-horizon rises to 61.2% from K3's 55.95%.
Game engines and Three.js work
Unity, Unreal Engine and Godot are supported, alongside web search and 3D generation; front-end output shows tighter layout, type and motion.
Scripted 3D and CAD
BenchCAD with tools reaches 92.7%, and it writes Blender or CAD scripts to shape assets, swap parts and adjust geometry, materials and layout — no MCP connectors required.
End-to-end finance tasks
OfficeQA Pro v2 scores 64.4; versus K3 it adds 7.2 points on Financial Modeling, 11.6 on Financial Research and about 4.2x on Business Operation.
Accuracy under pressure
On the internal Kimi Hallu Bench Hard, claim-level hallucination falls from 23.4% on K3 to 8.3%, and answers with at least one factual slip drop to 42.4% of K3's rate.
Cyber safeguards in place
Cyber data was removed from mid-training and RL; the shipped model passes 81.0% of xdev-bench-hard at pass@3, versus 90.5% for the checkpoint without those filters.
Kimi K3.1 Questions, Answered
Answers about Kimi K3.1, the faster build, open-weight timing and the places you can run it today.
What is Kimi K3.1?
It is the first release in Kimi's K3.1 series: K3 with much heavier post-training, offered as an openly released three-trillion-scale model with space for a million tokens.
How does it improve on Kimi K3?
Coding, agent and knowledge tests all move up — 61.2% versus 55.95% on Kimi Code Bench 3.0 Long-horizon, while hallucination drops from 23.4% to 8.3%.
When will the open weights be released?
Full weights are planned for release by Nov 4, 2026; hosted access through Kimi's products is available now.
Where can I use it right now?
Use it on kimi.ai and the mobile apps, in Kimi Work 3.1.0 or newer, in Kimi Code, and through the Kimi API, which offers Zero Data Retention.
Can I get faster responses?
Yes — the rapid variant replies 2-3x faster and can be called as kimi-k3.1-pro-highspeed on the Kimi API Platform.
How does it compare with closed frontier models?
It lands near the top closed models: 65.7% on Terminal-Bench 4.0 (Claude Opus 5.5 at 66.4%, GPT-6 Astra at 58.2%) and 92.7% on BenchCAD with tools.
Start Building with Kimi K3.1
Use kimi k3.1 on Kimi.ai or through the API, switch to the faster build when latency matters, and watch for full weights by Nov 4, 2026.