Can DeepSeek R1 Distill Qwen 32B run on the RX 7900 XT?
Yes. At 8K context the best fit is Q2_K, using about 14.2 GB of the 19.0 GB usable on the RX 7900 XT. Estimated speed ceiling: ~65 tokens/sec.
Better quality with offload: 4-bit (Q4_K_M)
Runs with offload 4-bit (Q4_K_M) at 8K context: 55 of 64 layers on the RX 7900 XT (about 18.3 GB) and the other 9 in system RAM (about 2.9 GB). Needs a PC with at least 32 GB of system RAM. Estimated speed: 6.6–11 tokens/sec (slow but usable), versus a bandwidth ceiling of ~40 tok/s if it fit fully in VRAM.
llama-server -hf bartowski/DeepSeek-R1-Distill-Qwen-32B-GGUF:Q4_K_M -c 8192 -ngl 55
Estimate, not a benchmark. The command downloads the GGUF from Hugging Face. The speed range spans dual-channel DDR4-3200 to DDR5-6000 system RAM; your CPU, RAM speed and llama.cpp version move it. If llama.cpp runs out of GPU memory, lower -ngl by 2-3; if VRAM is left over, raise it for more speed. Recent llama.cpp builds can also pick the split automatically. How we estimate offload.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 13.2 GB Fits | 14.2 GB Fits | 20.2 GB No | 44.2 GB No | 65 |
| 3-bit (Q3_K_M) • | 16.5 GB Fits | 17.5 GB Tight | 23.5 GB No | 47.5 GB No | 50 |
| 4-bit (Q4_K_M) • | 20.2 GB No | 21.2 GB No | 27.2 GB No | 51.2 GB No | — |
| 5-bit (Q5_K_M) • | 23.4 GB No | 24.4 GB No | 30.4 GB No | 54.4 GB No | — |
| 6-bit (Q6_K) • | 26.7 GB No | 27.7 GB No | 33.7 GB No | 57.7 GB No | — |
| 8-bit (Q8_0) • | 34.1 GB No | 35.1 GB No | 41.1 GB No | 65.1 GB No | — |
| 16-bit (F16/BF16) • | 62.7 GB No | 63.7 GB No | 69.7 GB No | 93.7 GB No | — |
• real GGUF file size from bartowski/DeepSeek-R1-Distill-Qwen-32B-GGUF. Speed is a bandwidth ceiling (800 GB/s), not a benchmark. Methodology.
What fits instead
- Cheapest listed GPU that runs DeepSeek R1 Distill Qwen 32B: RTX 5090 (32 GB, launch MSRP $1,999) at Q5_K_M, fully in VRAM at 8K context.
- Largest popular model that fits the RX 7900 XT: MiniMax H3 Image PE (27B) at Q4_K_M, about 16.8 GB at 8K context.
"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.
More
DeepSeek R1 Distill Qwen 32B on every GPU and Mac · Every model that runs on the RX 7900 XT · Other models that fit 24 GB VRAM · Try other settings in the calculator
Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs
Data: Hugging Face Hub (deepseek-ai/DeepSeek-R1-Distill-Qwen-32B), AMD specs. Model data last checked 2026-11-18; refreshed hourly.