TITAN RTX
NVIDIA
24GB VRAM · 672 GB/s bandwidth · 32.6 FP16 TFLOPS · 280W TDP
The TITAN RTX has 24GB of VRAM with 672 GB/s memory bandwidth and 32.6 TFLOPS FP16 compute. At Q4 quantization, it can comfortably run Gemma 3 4B (194 tok/s), Qwen 2.5 7B (102 tok/s), Llama 3.1 8B (97 tok/s). Models larger than ~41B parameters won't fit even at Q4. Electricity cost is approximately $30/month at 280W TDP.
What LLMs can you run?
| Model | Params | Q4 Weight | Fit | Decode |
|---|---|---|---|---|
| Gemma 3 4B | 4.0B | 2 GB | comfortable | 194 tok/s |
| Qwen 2.5 7B | 7.6B | 4 GB | comfortable | 102 tok/s |
| Llama 3.1 8B | 8.0B | 4 GB | comfortable | 97 tok/s |
| Mistral Small 24B | 24.0B | 12 GB | comfortable | 32 tok/s |
| Gemma 3 27B | 27.4B | 14 GB | tight | 28 tok/s |
| Qwen 2.5 Coder 32B | 32.5B | 16 GB | tight | 23 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 |
|---|---|---|---|---|
| Quadro RTX 6000 | 24GB | 672 GB/s | 32.6 | 260W |
| Tesla P10 | 24GB | 694 GB/s | 0.2 | 150W |
| Quadro RTX 6000 Passive | 24GB | 624 GB/s | 29.9 | 260W |
| A10 PCIe | 24GB | 600 GB/s | 31.2 | 150W |
| A10G | 24GB | 600 GB/s | 31.5 | 150W |
Compare with another GPU
Select another GPU to compare specs and model performance side by side.