Can the Apple M4 (16 GB) run DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic?
DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic·9.4B·Apple M4 (16 GB)·16 GB unified memory
Yes. The Apple M4 (16 GB) runs DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic fully in unified memory at Q6_K (about 8.9 GB of 11 GB), at roughly 8.5 tokens/s.
- best quant
- Q6_K
- needs
- 8.9 GB
- usable
- 11 GB
- est. speed
- ~8.5 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic page →
Every quantisation of DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic on the Apple M4 (16 GB)
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 5.3 GB | fits | ~21 tok/s |
| Q3_K_M | 6.1 GB | fits | ~14 tok/s |
| Q4_K_M | 7.2 GB | fits | ~12 tok/s |
| Q5_K_M | 8.0 GB | fits | ~9.9 tok/s |
| Q6_K pick | 8.9 GB | fits | ~8.5 tok/s |
| Q8_0 | 11 GB | too big | — |
| FP16 / BF16 | 20 GB | too big | — |
Speed is a bandwidth-based estimate for generating text (120 GB/s); prompt processing and long contexts are slower.
Measured speed on the Apple M4 (16 GB)
No community reports for this pair yet.
Report your own speed →Run DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic on the Apple M4 (16 GB)
Using Q6_K, the best fit for Apple M4 (16 GB).
pip install mlx-lm
mlx_lm.generate --model laewr/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-MLX-4bit --prompt "Hello"Use `mlx_lm.chat` with the same --model for an interactive chat.
mlx_lm.server --model laewr/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-MLX-4bit --port 8080API on http://localhost:8080/v1.
ollama run hf.co/mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF:Q6_K llama-server -hf mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF:Q6_Kbrew install llama.cpp ships with Metal. Downloads the GGUF on first run.
lms get mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF@Q6_KThen load it from the LM Studio app or with `lms load`.
from transformers import pipeline
pipe = pipeline("text-generation", model="xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic", dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "Explain self-hosting in one paragraph."}]
out = pipe(messages, max_new_tokens=512)
print(out[0]["generated_text"][-1]["content"])Chat models that fit the Apple M4 (16 GB) comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 8.2B | Q8_0 | ≈10 GBFits in memory | ~7.6 tok/s | |
| 8B | Q8_0 | ≈9.8 GBFits in memory | ~7.7 tok/s | |
| 9.2B | Q5_K_M | ≈9.6 GBFits in memory | ~10 tok/s | |
| 8B | Q8_0 | ≈9.8 GBFits in memory | ~7.7 tok/s | |
| 7.6B | Q8_0 | ≈8.9 GBFits in memory | ~8.2 tok/s | |
| 9.7B | Q6_K | ≈9.0 GBFits in memory | ~8.3 tok/s |
Other GPUs that run DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic
Frequently asked questions
- Can the Apple M4 (16 GB) run DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic?
- Yes. The Apple M4 (16 GB) runs DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic fully in unified memory at Q6_K (about 8.9 GB of 11 GB), at roughly 8.5 tokens/s.
- How fast does DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic run on the Apple M4 (16 GB)?
- An estimate from memory bandwidth gives about 8.5 tokens/s at Q6_K.
- Which quantisation of DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic should I use on the Apple M4 (16 GB)?
- Q6_K: the largest one that fits. Expected quality: near-lossless compared with the full-precision weights.
Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.