UTC

Kimi K3

Moonshot AI · open weights · served by 130 providers · tool calling, reasoning, image input, video input

moonshotai/kimi-k3

Weights
Open
Released
2026-07-16
Knowledge cutoff
Not stated
Context window
1.048576M
Deprecation dates

When each platform deprecates Kimi K3, as LiteLLM records it. Plan the move before the date: after it, the platform may stop serving the model or stop updating it.

  • GitHub Copilotdeprecates 2026-12-14github_copilot/kimi-k3
Every upcoming deprecation →
Hugging Face repository

moonshotai/Kimi-K3 at commit f831ab6681, last changed 2026-09-02

Licence
kimi-k3
Access
Open download
Library
transformers
Formats
.safetensors
  • Loading it runs code from the repository (trust_remote_code). Read that code, and pin the commit.
  • No model signature file is published with the weights.

What the Hub's scanners report

32 weight files, 32 with a scan result. No scanner reports an issue with them.

The model at this commit, its 32 weight files with their SHA-256.

Facts from the Hugging Face Hub and its file scanners (Protect AI, Cisco, Hugging Face picklescan, VirusTotal, JFrog), as of 2026-11-18. ShieldMap does not copy model cards.

Who serves it
  • 302.AIkimi-k3
  • Abacusmoonshotai/Kimi-K3
  • AbliteratedLLMkimi-k3
  • AbliteratedLLMkimi-k3-fast
  • Adzuki AIkimi-k3
  • ai&moonshotai/kimi-k3
  • AIHubMixkimi-k3
  • AI/ML APImoonshot/kimi-k3
  • ainetcafeKimi-K3
  • Alibabakimi-k3
  • Alibaba (China)kimi-k3
  • Amazon Bedrockglobal.moonshotai.kimi-k3
  • Amazon Bedrockus.moonshotai.kimi-k3
  • Arceemoonshotai/kimi-k3beta
  • Basetenmoonshotai/Kimi-K3
  • Berget.AImoonshotai/Kimi-K3
  • Boundlesskimi-k3
  • cheapaimodelkimi-k3
  • CheapestTokenkimi-k3
  • Chutesmoonshotai/Kimi-K3-TEE
  • ClinePasscline-pass/kimi-k3
  • Cloudflare AI Gatewaymoonshotai/kimi-k3
  • CommandCodemoonshotai/Kimi-K3
  • Cortecskimi-k3
106 more providers

CrossModel (moonshot/kimi-k3), Deep Infra (moonshotai/Kimi-K3), DigitalOcean (kimi-k3), DotHoosh (kimi-k3), Doubleword (moonshotai/kimi-k3), Eden AI (moonshot/kimi-k3), EmpirioLabs AI (kimi-k3), engy (kimi-k3), Fireworks AI (accounts/fireworks/models/kimi-k3), Fireworks AI (accounts/fireworks/routers/kimi-fast-latest), Fireworks AI (accounts/fireworks/routers/kimi-k3-fast), Fireworks AI (accounts/fireworks/routers/kimi-k3-spot), Fireworks AI (accounts/fireworks/routers/kimi-latest), FrogBot (kimi-k3), GitHub Copilot (kimi-k3), GreenPT (kimi-k3), Hostinger Router (kimi-k3), Hugging Face (moonshotai/Kimi-K3), Charm Hyper (kimi-k3), Impossibl (moonshotai/kimi-k3), Inco (kimi-k3:fast), Inferris (kimi-k3), IteraCompute (moonshotai/kimi-k3), Jalapeno Cloud (Kimi-K3), JZSTOKEN (kimi-k3), Kenari (kimi-k3), Kilo Gateway (moonshotai/kimi-k3), Kilo Gateway (moonshotai/kimi-k3-fast), Kimi For Coding (kimi.com) (k3), Kimi For Coding (kimi.com) (k3-256k), Kimi For Coding (kimi.ai) (k3), Kimi For Coding (kimi.ai) (k3-256k), KoRouter (kimi-k3), Liquid Inference (kimi-k3), DevPass (LLM Gateway) (kimi-k3), DevPass (LLM Gateway) (kimi-k3-fast), LLM Gateway (alibaba/kimi-k3), LLM Gateway (aws-bedrock/kimi-k3), LLM Gateway (baidu/kimi-k3), LLM Gateway (canopywave/kimi-k3), LLM Gateway (fireworks/kimi-k3), LLM Gateway (fireworks/kimi-k3-fast), LLM Gateway (moonshot/kimi-k3), LLM Gateway (novita/kimi-k3), LLM Gateway (runpod/kimi-k3), LLM Gateway (runware/kimi-k3), LLM Gateway (scx-ai-gp/kimi-k3), LLM Gateway (tencent/kimi-k3), LLM Gateway (together-ai/kimi-k3), Melious (kimi-k3), Merge Gateway (moonshot/kimi-k3), Moark (kimi-k3), Modal (moonshotai/Kimi-K3), Moonshot AI (kimi-k3), Moonshot AI (China) (kimi-k3), Morph (morph-kimik3), Morph (morph-kimik3-fast), NanoGPT (TEE/kimi-k3), NanoGPT (moonshotai/kimi-k3), NEAR AI Cloud (moonshotai/kimi-k3), Nebius Token Factory (moonshotai/Kimi-K3), Neon (kimi-k3), Novita AI (moonshotai/kimi-k3), Nvidia (moonshotai/kimi-k3), Ofox (moonshotai/kimi-k3), Ollama Cloud (kimi-k3), OpenCode (kimi-k3), OpenCode Go (kimi-k3), OpenRouter (moonshotai/kimi-k3), Opper (kimi-k3), OrcaRouter (kimi/kimi-k3), Perplexity Agent (moonshot-ai/kimi-k3), Perplexity Agent (perplexity/kimi-k3), Phala (moonshotai/kimi-k3), Pioneer (moonshotai/Kimi-K3), Pioneer (moonshotai/Kimi-K3-Fast), Prime Inference (moonshotai/kimi-k3), Requesty (kimi-k3), Requesty (kimi-k3@eu), SandBase (moonshotai/kimi-k3), SCNet Token Plan (Kimi-K3), SCX.ai (Kimi-K3), SenseNova (China) (kimi-k3), SiliconFlow (moonshotai/Kimi-K3), Standard Compute (moonshotai/kimi-k3), SteadyRouter (kimi-k3), Synthetic (hf:moonshotai/Kimi-K3), Synthorai (kimi-k3), Tencent Token Plan (kimi-k3), TensorX (moonshotai/kimi-k3), Together AI (moonshotai/Kimi-K3), TokenGo (moonshotai/kimi-k3), UplinkSwitch (kimi-k3), Vancine (kimi-k3), Venice AI (kimi-k3), Venice AI (kimi-k3-fast-api), Vercel AI Gateway (moonshotai/kimi-k3), Vercel AI Gateway (moonshotai/kimi-k3-fast), Vivgrid (kimi-k3), Volcengine Ark Coding Plan (kimi-k3), Wafer (Kimi-K3), Wallaby (moonshotai/kimi-k3), Wally (kimi-k3), Watt Cloud (Kimi-K3), Zenifra (zenifra/kimi-k3), ZenMux (moonshotai/kimi-k3)

Running it yourself

Open weights are loaded and served by code that has its own advisories. These inference servers and model loaders are on ShieldMap; load weights only from the publisher's own repository, and prefer safetensors over formats that can run code.

How supply-chain attacks on models work →

Sources: models.dev (MIT) and the LiteLLM model catalogue (MIT). Deprecation dates are as each platform published them and as LiteLLM records them; check the platform before you rely on one. Data as of 2026-11-22.