Radeon PRO W6800
AMD
32GB VRAM · 512 GB/s bandwidth · 35.7 FP16 TFLOPS · 250W TDP
The Radeon PRO W6800 has 32GB of VRAM with 512 GB/s memory bandwidth and 35.7 TFLOPS FP16 compute. At Q4 quantization, it can comfortably run Gemma 3 4B (115 tok/s), Qwen 2.5 7B (60 tok/s), Llama 3.1 8B (57 tok/s). Models larger than ~54B parameters won't fit even at Q4. Electricity cost is approximately $27/month at 250W TDP.
What LLMs can you run?
| Model | Params | Q4 Weight | Fit | Decode |
|---|---|---|---|---|
| Gemma 3 4B | 4.0B | 2 GB | comfortable | 115 tok/s |
| Qwen 2.5 7B | 7.6B | 4 GB | comfortable | 60 tok/s |
| Llama 3.1 8B | 8.0B | 4 GB | comfortable | 57 tok/s |
| Mistral Small 24B | 24.0B | 12 GB | tight | 19 tok/s |
| Gemma 3 27B | 27.4B | 14 GB | tight | 16 tok/s |
| Qwen 2.5 Coder 32B | 32.5B | 16 GB | tight | 14 tok/s |
| Llama 3.3 70B | 70.6B | 35 GB | won't fit | |
| Qwen 2.5 72B | 72.7B | 36 GB | won't fit | |
| Llama 3.1 405B | 405B | 202 GB | won't fit | |
| DeepSeek R1 671B | 671B | 336 GB | won't fit |
Similar GPUs
| GPU | VRAM | BW | TFLOPS | TDP |
|---|---|---|---|---|
| Radeon PRO V620 | 32GB | 512 GB/s | 40.5 | 300W |
| Radeon AI PRO 9600D | 32GB | 576 GB/s | 49.6 | 150W |
| Radeon PRO W7800 | 32GB | 576 GB/s | 90.5 | 260W |
| RTX 5000 Ada Generation | 32GB | 576 GB/s | 65.3 | 250W |
| Radeon AI PRO R9700 | 32GB | 644 GB/s | 95.7 | 300W |
Compare with another GPU
Select another GPU to compare specs and model performance side by side.