Can I run Qwen3.6 35B-A3B on the Apple M4 Pro?
Runs greatQ4_K_M fits in unified memory (23.0 of 33.6 GB usable by the GPU): Runs great, est. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%).
48 GB unified memory · 273.0 GB/s · FP16 8.1 TFLOPS. Assumes 8k context and an f16 KV cache. Verdicts are shown for each memory size.
Verdict by memory size
The page uses the 48 GB configuration; here is every size.
- 24 GBWon't runTry IQ4_XS: Runs slowlyest. 64.8 tok/s (45.4–84.2, theoretical estimate ±30%)Q2_K (heavy quality loss): Runs great
- 48 GBRuns greatest. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%)(used on this page)
- 64 GBRuns greatest. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%)
Every quant of Qwen3.6 35B-A3B on the Apple M4 Pro
| Quant | File size | Verdict | Speed | Memory | Runs as | Notes |
|---|---|---|---|---|---|---|
| Q2_K | 13.51 GB | Runs great | est. 83.5 tok/s66.8–100.2calibrated estimate ±20% | 14.3 / 33.6 GB | Unified memory | — |
| IQ4_XS | 19.70 GB | Runs great | est. 75.1 tok/s60.1–90.1calibrated estimate ±20% | 20.4 / 33.6 GB | Unified memory | — |
| Q4_K_Mbaseline | 22.29 GB | Runs great | est. 68.8 tok/s55.0–82.5calibrated estimate ±20% | 23.0 / 33.6 GB | Unified memory | — |
| Q8_0 | 37.81 GB | Runs slowly | est. 35.3 tok/s24.7–45.9theoretical estimate ±30% | 38.0 GB RAM | CPU only | CPU inference — capped at Runs slowly |
Models with a thinking mode spend extra tokens thinking, so their speed thresholds are 1.5× stricter (30 / 12 / 3 tok/s). With thinking turned off, a model may do one grade better.
Where the memory goes at Q4_K_M
- Weights
- 22.3 GB
- KV cache
- 0.2 GB
- Compute buffer
- 0.6 GB
- Free
- 10.6 GB
- Not usable by the GPU
- 14.4 GB
Context length vs. KV cache
| Context | KV f16 | KV q8_0 | KV q4_0 |
|---|---|---|---|
| 4k | 22.9 GBRuns great est. 70.7 tok/s56.6–84.9calibrated estimate ±20% | 22.9 GBRuns great est. 71.7 tok/s57.3–86.0calibrated estimate ±20% | 22.9 GBRuns great est. 72.2 tok/s57.8–86.6calibrated estimate ±20% |
| 8k | 23.0 GBRuns great est. 68.8 tok/s55.0–82.5calibrated estimate ±20% | 23.0 GBRuns great est. 70.6 tok/s56.5–84.7calibrated estimate ±20% | 22.9 GBRuns great est. 71.6 tok/s57.3–85.9calibrated estimate ±20% |
| 16k | 23.3 GBRuns great est. 65.1 tok/s52.1–78.2calibrated estimate ±20% | 23.1 GBRuns great est. 68.5 tok/s54.8–82.2calibrated estimate ±20% | 23.0 GBRuns great est. 70.4 tok/s56.3–84.5calibrated estimate ±20% |
| 32k | 23.8 GBRuns great est. 58.9 tok/s47.1–70.7calibrated estimate ±20% | 23.4 GBRuns great est. 64.7 tok/s51.7–77.6calibrated estimate ±20% | 23.3 GBRuns great est. 68.2 tok/s54.6–81.9calibrated estimate ±20% |
| 64k | 24.7 GBRuns great est. 49.5 tok/s39.6–59.4calibrated estimate ±20% | 24.1 GBRuns great est. 58.1 tok/s46.5–69.8calibrated estimate ±20% | 23.8 GBRuns great est. 64.2 tok/s51.3–77.0calibrated estimate ±20% |
| 128k | 26.7 GBRuns great est. 37.5 tok/s30.0–45.0calibrated estimate ±20% | 25.4 GBRuns great est. 48.4 tok/s38.7–58.1calibrated estimate ±20% | 24.8 GBRuns great est. 57.4 tok/s45.9–68.9calibrated estimate ±20% |
| 256k | 30.6 GBRuns well est. 25.2 tok/s20.2–30.3calibrated estimate ±20% | 28.1 GBRuns great est. 36.3 tok/s29.0–43.5calibrated estimate ±20% | 26.7 GBRuns great est. 47.4 tok/s37.9–56.8calibrated estimate ±20% |
S = Runs great · A = Runs well · B = Runs slowly · F = Won't run
At Q4_K_M the verdict stays Runs great up to 128k context with an f16 KV cache, and up to 256k with q8_0.
Estimated speed
- Generation
- est. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%)
- Prompt processing
- about 340 tok/s (204–476, rough estimate ±40%)
Measured on this exact combination
No public measurement for this exact combination yet; the numbers above are estimates.
If this is not enough
A larger model that also runs well on the Apple M4 Pro
Runs greatK2-Horizon MoVA 36B-A4B: Runs great, est. 36.8 tok/s (25.8–47.8, theoretical estimate ±30%)
How to run it
Commands for the Q4_K_M file. Both tools download from Hugging Face on first run.
File: Qwen_Qwen3.6-35B-A3B-Q4_K_M.gguf (22.29 GB) from bartowski/Qwen_Qwen3.6-35B-A3B-GGUF.
llama-server -hf bartowski/Qwen_Qwen3.6-35B-A3B-GGUF:Q4_K_M -c 8192 -ngl all -fa on- -ngl all loads every layer on the GPU.
- -fa on enables flash attention (needed for KV cache quantization).
- --cache-type-k q8_0 --cache-type-v q8_0 shrinks the KV cache to 0.09 GB at 8k context.
Terminal:
OLLAMA_CONTEXT_LENGTH=8192 ollama serve
ollama run hf.co/bartowski/Qwen_Qwen3.6-35B-A3B-GGUF:Q4_K_M- Ollama defaults to 4k context on GPUs under 24 GB. Set OLLAMA_CONTEXT_LENGTH when starting the server, or /set parameter num_ctx inside the chat.
/set parameter num_ctx 8192
Related pages
Guide: Running local LLMs on a Mac (M4, M5): how much unified memory?
Other models on the Apple M4 Pro
Qwen3.6 35B-A3B on other hardware
Frequently asked questions
Can I run Qwen3.6 35B-A3B on the Apple M4 Pro?
Q4_K_M fits in unified memory (23.0 of 33.6 GB usable by the GPU): Runs great, est. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%). Q4_K_M needs 23.0 GB at 8k context; the GPU can use 33.6 GB of the unified memory.
How much context can Qwen3.6 35B-A3B use on the Apple M4 Pro?
At Q4_K_M the verdict stays Runs great up to 128k context with an f16 KV cache (2.7 GB of KV) and up to 256k with a q8_0 KV cache (2.9 GB). The model supports up to 256k.
Which quant should I use, and how fast is it?
Q4_K_M (22.29 GB) is the recommended balance: Runs great, est. 68.8 tok/s (55.0–82.5, calibrated estimate ±20%). It is also the largest tracked file that gets this verdict.
Speeds are estimates from memory bandwidth, calibrated against public benchmarks, and each one comes with an error band and a confidence label. Real results vary with drivers, backend, context length and thermals.