RTX 6000 Ada Generation
NVIDIA
48GB VRAM · 960 GB/s bandwidth · 91.1 FP16 TFLOPS · 300W TDP
The RTX 6000 Ada Generation has 48GB of VRAM with 960 GB/s memory bandwidth and 91.1 TFLOPS FP16 compute. At Q4 quantization, it can comfortably run Gemma 3 4B (312 tok/s), Qwen 2.5 7B (164 tok/s), Llama 3.1 8B (155 tok/s). Models larger than ~82B parameters won't fit even at Q4. Electricity cost is approximately $32/month at 300W TDP.
What LLMs can you run?
| Model | Params | Q4 Weight | Fit | Decode |
|---|---|---|---|---|
| Gemma 3 4B | 4.0B | 2 GB | comfortable | 312 tok/s |
| Qwen 2.5 7B | 7.6B | 4 GB | comfortable | 164 tok/s |
| Llama 3.1 8B | 8.0B | 4 GB | comfortable | 155 tok/s |
| Mistral Small 24B | 24.0B | 12 GB | comfortable | 52 tok/s |
| Gemma 3 27B | 27.4B | 14 GB | comfortable | 45 tok/s |
| Qwen 2.5 Coder 32B | 32.5B | 16 GB | comfortable | 38 tok/s |
| Llama 3.3 70B | 70.6B | 35 GB | tight | 17 tok/s |
| Qwen 2.5 72B | 72.7B | 36 GB | tight | 17 tok/s |
| 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 W7800 48 GB | 48GB | 864 GB/s | 90.5 | 281W |
| Radeon PRO W7900 | 48GB | 864 GB/s | 122.6 | 295W |
| Radeon PRO W7900D | 48GB | 864 GB/s | 106.0 | 295W |
| RTX 5880 Ada Generation | 48GB | 864 GB/s | 69.3 | 285W |
| RTX A6000 | 48GB | 768 GB/s | 38.7 | 300W |
Compare with another GPU
Select another GPU to compare specs and model performance side by side.