No — not on this device
Kimi K2.6 1T-A32B at Q4_K_M needs 578.6 GB against 14.4 GB usable, and the shortfall of 564.2 GB is more than 256 GB of system RAM can cover at a tolerable speed. A smaller sibling or a lower quantisation is the honest answer here.
Does not fit
8K context
Q4_K_M · 577.5 GB
Modified MIT
Released 20 Apr 2026
Not in the Ollama library
The open coding-agent benchmark leader of spring 2026 (80.2 SWE-bench). A 512 GB Mac Studio pair, or a ceiling.
The VRAM budget
weights 577.5 GB
Weights 577.5 GB
KV cache @ 8K 0.54 GB
Runtime overhead 0.6 GB
Over budget 564.2 GB past 14.4 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 1016.2 GB | 1017.4 GB | — | ~0.4 | −0.1% ppl | 1003.0 GB over |
| Q6_K | 784.3 GB | 785.4 GB | — | ~0.6 | −0.4% ppl | 771.0 GB over |
| Q5_K_M | 677.9 GB | 679.0 GB | — | ~0.7 | −0.8% ppl | 664.6 GB over |
| Q4_K_M | 577.5 GB | 578.6 GB | — | ~0.8 | −1.9% ppl | 564.2 GB over |
| Q3_K_M | 467.5 GB | 468.6 GB | — | ~1.0 | −5.4% ppl | 454.2 GB over |
| Q2_K | 400.5 GB | 401.7 GB | — | ~1.1 | −15% ppl | 387.3 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. This model uses multi-head latent attention, so its cache is a compressed latent rather than full K and V.
How to run it
$ llama-server \
-hf moonshotai/Kimi-K2.6:Q4_K_M \
-c 8192 -ngl 1
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
01Download is 577.5 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03If it falls back to CPU silently, drop the context first, then step down a quantisation.