NVIDIA H200 SXM 141GB meets the minimum VRAM requirement for Q4 inference of deepseek-ai/DeepSeek-OCR. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
NVIDIA H200 SXM 141GB can run deepseek-ai/DeepSeek-OCR with Q4 quantization. At approximately 853 tokens/second, you can expect Excellent speed - conversational response times under 1 second.
You have 139GB headroom, which is sufficient for system overhead and smooth operation.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 2GB | 141GB | 853.31 tok/s | ✅ Fits comfortably |
| Q8 | 4GB | 141GB | 576.97 tok/s | ✅ Fits comfortably |
| FP16 | 7GB | 141GB | 302.69 tok/s | ✅ Fits comfortably |