# Can the RTX 4060 run DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic?

> Yes. The RTX 4060 runs DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic fully in VRAM at Q4_K_M (about 7.2 GB of 8.0 GB), at roughly 26 tokens/s.

GeForce RTX 4060: 8 GB (8 GB usable), 272 GB/s. Estimates at an 8K context.

## Every quantisation

| Quant | Quality | ≈ GB | Fit | Est. tok/s |
|---|---|---|---|---|
| Q2_K | Heavy loss | 5.33 | fits | 48 |
| Q3_K_M | Noticeable loss | 6.13 | fits | 33 |
| Q4_K_M (pick) | Small loss | 7.15 | fits | 26 |
| Q5_K_M | Near-lossless | 8 | tight | 22 |
| Q6_K | Near-lossless | 8.89 | offload | 14 |
| Q8_0 | Lossless | 11.07 | offload | 7.6 |
| F16 | Lossless | 19.5 | offload | 2.6 |

## Run it

Q4_K_M straight from Hugging Face (ollama):

```bash
ollama run hf.co/mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF:Q4_K_M
```

OpenAI-compatible server with a web UI (llama.cpp):

```bash
llama-server -hf mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF:Q4_K_M
```

Download with the lms CLI (lmstudio):

```bash
lms get mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-GGUF@Q4_K_M
```

Serve an OpenAI-compatible API (vllm):

```bash
vllm serve xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic
```

Docker Compose (NVIDIA) (vllm):

```yaml
services:
  vllm:
    image: vllm/vllm-openai:latest
    command: xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic
    ports:
      - "8000:8000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
```

Serve an OpenAI-compatible API (sglang):

```bash
sglang serve xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic --host 0.0.0.0 --port 30000
```

Docker Compose (NVIDIA) (sglang):

```yaml
services:
  sglang:
    image: lmsysorg/sglang:latest
    command: sglang serve xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic --host 0.0.0.0 --port 30000
    ports:
      - "30000:30000"
    ipc: host
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
```

Chat with the pipeline API (transformers):

```python
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 comfortably

- [Qwen/Qwen3-8B](https://selfhost.directory/models/qwen/qwen3-8b/on/rtx-4060): Q5_K_M ≈7.35 GB, est. 26 tok/s
- [meta-llama/Llama-3.1-8B-Instruct](https://selfhost.directory/models/meta-llama/llama-3.1-8b-instruct/on/rtx-4060): Q5_K_M ≈7.11 GB, est. 26 tok/s
- [meta-llama/Meta-Llama-3-8B-Instruct](https://selfhost.directory/models/meta-llama/meta-llama-3-8b-instruct/on/rtx-4060): Q5_K_M ≈7.11 GB, est. 26 tok/s
- [Qwen/Qwen2.5-7B-Instruct](https://selfhost.directory/models/qwen/qwen2.5-7b-instruct/on/rtx-4060): Q6_K ≈7.05 GB, est. 24 tok/s
- [ornith-ai/Ornith-1.5-9B](https://selfhost.directory/models/ornith-ai/ornith-1.5-9b/on/rtx-4060): Q4_K_M ≈7.28 GB, est. 26 tok/s
- [ibm-granite/granite-4.2-8b](https://selfhost.directory/models/ibm-granite/granite-4.2-8b/on/rtx-4060): Q4_K_M ≈6.98 GB, est. 28 tok/s

## Other GPUs that run it

- [Apple M4 (16 GB)](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/apple-m4-16gb): Q6_K, est. 8.5 tok/s
- [Radeon RX 6700 XT](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/rx-6700-xt): Q8_0, est. 21 tok/s
- [GeForce RTX 3060 12GB](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/rtx-3060): Q8_0, est. 20 tok/s
- [Arc B580](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/arc-b580): Q8_0, est. 25 tok/s
- [Tesla T4](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/tesla-t4): Q8_0, est. 18 tok/s
- [Arc A770 16GB](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/arc-a770): Q8_0, est. 31 tok/s
- [Radeon RX 7800 XT](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/rx-7800-xt): Q8_0, est. 34 tok/s
- [Tesla P40](https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/tesla-p40): F16, est. 10 tok/s

---
Model: https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic. GPU: https://selfhost.directory/models/gpus/rtx-4060. Source: https://selfhost.directory/models/xtools-at/deepseek-v4-pro-qwen3.5-9b-uncensored-heretic/on/rtx-4060