NVIDIA H200 SXM 141GB meets the minimum VRAM requirement for Q4 inference of ai-forever/ruGPT-3.5-13B. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
NVIDIA H200 SXM 141GB can run ai-forever/ruGPT-3.5-13B with Q4 quantization. At approximately 473 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 134GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 7GB | 141GB | 472.55 tok/s | ✅ Fits comfortably |
| Q8 | 13GB | 141GB | 348.92 tok/s | ✅ Fits comfortably |
| FP16 | 27GB | 141GB | 195.84 tok/s | ✅ Fits comfortably |