Put Kimi K3.1 Pro to Work
Hand over code, 3D models or research briefs to K3.1 Pro and inspect every result in one shared workspace
None

None

None

Long Story Video Skill

Turn a brief into a 10s–10min piece — the agent drafts the script, prompts and shots in one pass.

Ads Video Skill

Ads Video Skill

Build product ads end to end, with scripts, prompts and footage generated together.

3D Science Explainer Video Skill

Turn scientific concepts into 3D explainer animations

AI Video Prompt Generator

Feedback

freeTrialImage.bannerPity

freeTrialImage.upgradeUnlock

  • ✓freeTrialImage.benefitHd
  • ✓freeTrialImage.benefitWatermark
  • ✓freeTrialImage.benefitUnlimited

kimi k3.1 pro

Explore Kimi K3.1 Pro — an open-weight 3T-class model with a 1M-token window, a Highspeed mode and standout coding, CAD and research results.

All Tools

Discover our comprehensive AI-powered animation toolkit

Why Builders Choose Kimi K3.1 Pro

The first open-weight release in the K3.1 line: a 3T-class model holding a million-token window, plus a Highspeed variant made for quicker exchanges.

  • Open Weights at 3T Scale, One Million Tokens
    The strongest open-weight launch in the K3.1 series to date, carrying a million-token window that spans codebases, 3D scenes and financial models.
  • A Highspeed Build Two to Three Times Quicker
    The Highspeed variant answers roughly 2-3x faster than the standard build, keeping back-and-forth sessions responsive. Both versions are live today.
  • Reinforcement Learning Tuned Across Many Fields
    Post-training was scaled far beyond base K3 — RL framework, long-context serving, sandboxes and rollout scheduling tuned together for steady returns.

Ways to Reach Kimi K3.1 Pro

Pick the entry point that fits you: browser and mobile apps, the Kimi Code and Kimi Work desktop tools, or the API and enterprise console.

Capabilities of Kimi K3.1 Pro

From coding and long-running autonomy to parametric CAD, finance research, accuracy gains and benchmark standing — here is the full range of K3.1 Pro.

Full-Project Coding and Game Engines

Quick patches in live repos and whole-project builds both land: Unity, Unreal Engine and Godot work, with tighter typography, animation and front-end detail.

Autonomy Over Long Runs and Self-Correction

In an internal walking-simulator trial of 167 missions and 420 reviews, K3.1 Pro traced faults and fixed 36 shoulder moment-arm errors before retraining its controller.

Parametric CAD and 3D Modeling

With tools it reaches 92.7% on BenchCAD, and it writes and runs Blender and CAD scripts on its own — no MCP connectors — for parts, buildings and custom furniture.

Auditable Finance and Research Workflows

OfficeQA Pro v2 lands at 64.4, alongside gains over K3 of 7.2 points in financial modeling, 11.6 in research and roughly 4.2x in simulated business operation.

Fewer Hallucinations on Real Work

Focused tuning brought claim-level hallucination on Kimi Hallu Bench Hard from 23.4% down to 8.3%, and answers carrying at least one factual error fell to 42.4% of the earlier rate.

Benchmark Position and Guarded Open Weights

Terminal-Bench 4.0 comes in at 65.7%, ahead of GPT-6 Astra at 58.2%, while cyber-related data removal moved xdev-bench-hard from 90.5% to 81.0%.

FAQ

Questions About Kimi K3.1 Pro

Answers on context size, Highspeed mode, weight release timing, benchmark comparisons, finance accuracy and enterprise access.

1

How much context can Kimi K3.1 Pro hold?

A 1-million-token window. It opens the K3.1 series and stands as the strongest open-weight 3T-class model released so far.

2

Standard vs Highspeed: what actually changes?

Highspeed replies two to three times faster than the standard build, which suits rapid back-and-forth. Both versions are live on Kimi.ai, Kimi Work, Kimi Code and the API.

3

When do the full model weights arrive?

Full weights are scheduled for release by Nov 4, 2026, under published principles — among them keeping offensive cyber capability below the frontier.

4

How does it stack up against closed frontier models?

It matches or nears the strongest proprietary systems on coding, agentic and knowledge tests: 65.7% on Terminal-Bench 4.0 against 58.2% for GPT-6 Astra and 66.4% for Claude Opus 5.5, plus 92.7% on BenchCAD with tools.

5

Can finance teams rely on its numbers?

It posts 64.4 on OfficeQA Pro v2 and beats K3 in modeling and research, while claim-level hallucination on Kimi Hallu Bench Hard fell from 23.4% to 8.3% — that bench is a deliberate stress test, not an everyday error rate.

6

How do organizations enable access?

Kimi Enterprise admins can turn K3.1 Pro on for their org — it is off by default at launch. Kimi Business adds enterprise data privacy, member management and separate personal and org accounts.

Start Using Kimi K3.1 Pro Now

K3.1 Pro is available on Kimi.ai, the mobile apps, Kimi Work, Kimi Code and the Kimi API. Turn it on in Kimi Enterprise, or subscribe to Kimi Business for your team.