Hugging FaceFor long coding sessions and research work, with text, pictures and video read by one model.
- Rating
- 5.0
- Downloads
- 312K
- Parameters
- 2.8T, 104B active
- Context
- 1M
- Download
- 467 to 1562 GB
- Commercial use
- With conditions
Too big for this Mac
Its smallest size needs 716 GB of memory, more than M4 · 16 GB can give it.
About
It uses screenshots and visuals to help with game making, website front ends and CAD design.
Limitations
When your intent is unclear, it may make choices on your behalf.
The maker admits a noticeable gap in ease of use behind Claude Fable 5 and GPT 5.6 Sol.
Good to know
In a 48-hour run on its own, it designed, tuned and checked a computer chip using open-source tools.
It edited its own teaser video from 56 source clips.
Sources: moonshotai/Kimi-K3 model card, Kimi K3 Tech Blog: Open Frontier Intelligence. Every line comes from one of these.
Ratings
5.0 out of 5, from how people voted on its answers.
LM Arena ratings of October 28, 2026 (CC BY 4.0): people vote between two answers without knowing which model wrote which. The bar ends at 1488, the best any open model has. The Arena tests the full model; a small size on a Mac answers a little worse.
Good at
Licence
Kimi K3 License. Commercial use: with conditions. MIT-style license. A Model-as-a-Service business whose group revenue passes $20 million in any 12 months needs a separate agreement with Moonshot AI before commercial use. A commercial product or service with more than 100 million monthly active users or $20 million in monthly revenue must display "Kimi K3" on its interface. The maker's page has the full terms.
Kimi K3 does not run well on any Mac Outmake lists; its smallest size needs more memory than the largest configuration has. 0 of 59 Mac configurations run it well.
Compare with
The nearest models in size, judged for the same Mac.
- Qwen3.8 2.4TReasons through long tasksToo big for this Mac · 731 GBGet
- DeepSeek V4 ProBuilt for long, many-step tasks that need a lot of text held in view at once.Too big for this Mac · 850 GBGet
- Kimi K2.6Long coding jobs and agentsToo big for this Mac · 341 GBGet
- GLM 5.3Whole software projectsToo big for this Mac · 254 GBGet
Sizes
| Size | Download | Memory at 8K | Memory at 32K | Open |
|---|---|---|---|---|
| UD-IQ2_XXSSmaller | 712 GB | 712 GB | 713 GB | Open |
| UD-Q2_K_XLBalanced | 862 GB | 862 GB | 863 GB | Open |
| UD-Q4_K_XLBest | 1510 GB | 1510 GB | 1510 GB | Open |
All 9 sizes
Memory is what llama.cpp's own calculator reported for each size when the catalog was built, with the app's settings. Smaller, Balanced and Best are the three sizes the store names.
Run It Anywhere
Kimi K3 is an open model, so any app that runs GGUF files can run it. These take the UD-Q2_K_XL size from unsloth straight from Hugging Face.
- LM StudioMac, Windows, Linux
Open - JanMac, Windows, Linux
Open - MstyMac, Windows, Linux
Open - OllamaMac, Windows, Linux
ollama run hf.co/unsloth/Kimi-K3-GGUF:UD-Q2_K_XL - llama.cppMac, Windows, Linux
llama-server -hf unsloth/Kimi-K3-GGUF:UD-Q2_K_XL - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/unsloth/Kimi-K3-GGUF:UD-Q2_K_XL
On a PC Graphics Card
| Video memory | Largest size that fits | Download | Memory at 8K |
|---|---|---|---|
| 6 GB | None worth running | needs 862 GB | |
| 8 GB | None worth running | needs 862 GB | |
| 10 GB | None worth running | needs 862 GB | |
| 11 GB | None worth running | needs 862 GB | |
| 12 GB | None worth running | needs 862 GB | |
| 16 GB | None worth running | needs 862 GB | |
| 20 GB | None worth running | needs 862 GB | |
| 24 GB | None worth running | needs 862 GB | |
| 32 GB | None worth running | needs 862 GB | |
| 48 GB | None worth running | needs 862 GB |
The largest size, of those that keep three bits a weight or more, whose memory at an 8K chat is under 92% of the card's; the rest is the desktop's. Estimated from the files, the same way as for a Mac; speed on a PC is not estimated yet. A size that doesn't fit the card still runs, slower, with part of it in ordinary memory.
On every Mac
| Mac | Size | Fit | Memory | Writes | Reads |
|---|---|---|---|---|---|
| M6 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M6 · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M6 · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Pro · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Pro · 48 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Pro · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Max · 36 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Max · 48 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Max · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Max · 128 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Ultra · 96 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Ultra · 256 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M5 Ultra · 512 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Pro · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Pro · 48 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Pro · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Max · 36 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Max · 48 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Max · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Max · 96 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M4 Max · 128 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 · 8 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Pro · 18 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Pro · 36 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Max · 36 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Max · 48 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Max · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Max · 96 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Max · 128 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Ultra · 96 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Ultra · 256 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M3 Ultra · 512 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 · 8 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 · 24 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Pro · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Pro · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Max · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Max · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Max · 96 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Ultra · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Ultra · 128 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M2 Ultra · 192 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 · 8 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Pro · 16 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Pro · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Max · 32 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Max · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Ultra · 64 GB | UD-IQ2_XXS | Too big | 716 GB | ||
| M1 Ultra · 128 GB | UD-IQ2_XXS | Too big | 716 GB |
Writes and reads are words a second (a token is about three quarters of a word), 4K tokens into a chat, estimated unless marked measured. The size is the one the store offers that Mac. Constants fitted October 11, 2026; how, and how wrong.
Badge for the model card
Shows on Kimi K3's Hugging Face card which Macs run it well, linked to this page. It updates every morning.
Sizes and files are read from unsloth/Kimi-K3-GGUF on Hugging Face and checked every morning. The rating is the model's score on the LM Arena text leaderboard (CC BY 4.0), where people vote between two answers without knowing which model wrote which. Catalog of November 4, 2026.