Hugging FaceFor long coding jobs and agents that work on their own, and it can read pictures and video too.
- Rating
- 4.7
- Downloads
- 4.7K
- Parameters
- 1T, 32B active
- Context
- 256K
- Download
- 341 to 2054 GB
- Commercial use
- With conditions
Too big for this Mac
Its smallest size needs 351 GB of memory, more than M4 · 16 GB can give it.
About
Built for background agents that run around the clock, managing schedules and running code unattended. It has two ways to answer, a Thinking mode and an Instant mode.
Limitations
Chatting about video is experimental and works only through the official API for now.
Good to know
Built to split one task across up to 300 helper agents taking 4,000 coordinated steps.
Sources: moonshotai/Kimi-K2.6 model card. Every line comes from one of these.
Ratings
4.7 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
Modified MIT. Commercial use: with conditions. MIT-style license. A commercial product or service with more than 100 million monthly active users or $20 million in monthly revenue must display "Kimi K2.6" on its interface. The maker's page has the full terms.
Kimi K2.6 runs well from a Mac with an M3 Ultra and 512 GB of memory, at UD-Q2_K_XL, writing about 18 words (24 tokens) a second. 2 of 59 Mac configurations run it well.
Compare with
The nearest models in size, judged for the same Mac.
- GLM 5.3Whole software projectsToo big for this Mac · 254 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
- Qwen3.8 2.4TReasons through long tasksToo big for this Mac · 731 GBGet
- Qwen3.5 397BFor conversations that mix text and pictures, with room to read a long report in one go.Too big for this Mac · 124 GBGet
Sizes
| Size | Download | Memory at 8K | Memory at 32K | Open |
|---|---|---|---|---|
| UD-Q2_K_XLBalanced | 341 GB | 341 GB | 342 GB | Open |
| UD-Q4_K_XLBest | 585 GB | 584 GB | 585 GB | Open |
All 4 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 K2.6 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-K2.6-GGUF:UD-Q2_K_XL - llama.cppMac, Windows, Linux
llama-server -hf unsloth/Kimi-K2.6-GGUF:UD-Q2_K_XL - Docker Model RunnerMac, Windows, Linux
docker model run hf.co/unsloth/Kimi-K2.6-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 341 GB | |
| 8 GB | None worth running | needs 341 GB | |
| 10 GB | None worth running | needs 341 GB | |
| 11 GB | None worth running | needs 341 GB | |
| 12 GB | None worth running | needs 341 GB | |
| 16 GB | None worth running | needs 341 GB | |
| 20 GB | None worth running | needs 341 GB | |
| 24 GB | None worth running | needs 341 GB | |
| 32 GB | None worth running | needs 341 GB | |
| 48 GB | None worth running | needs 341 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-Q2_K_XL | Too big | 351 GB | ||
| M6 · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M6 · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Pro · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Pro · 48 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Pro · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Max · 36 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Max · 48 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Max · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Max · 128 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Ultra · 96 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Ultra · 256 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M5 Ultra · 512 GB | UD-Q2_K_XL | Runs great | 346 GB | 32 | 508 |
| M4 · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Pro · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Pro · 48 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Pro · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Max · 36 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Max · 48 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Max · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Max · 96 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M4 Max · 128 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 · 8 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Pro · 18 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Pro · 36 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Max · 36 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Max · 48 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Max · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Max · 96 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Max · 128 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Ultra · 96 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Ultra · 256 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M3 Ultra · 512 GB | UD-Q2_K_XL | Runs great | 346 GB | 18 | 141 |
| M2 · 8 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 · 24 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Pro · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Pro · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Max · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Max · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Max · 96 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Ultra · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Ultra · 128 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M2 Ultra · 192 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 · 8 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Pro · 16 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Pro · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Max · 32 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Max · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Ultra · 64 GB | UD-Q2_K_XL | Too big | 351 GB | ||
| M1 Ultra · 128 GB | UD-Q2_K_XL | Too big | 351 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 K2.6'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-K2.6-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.