NVIDIA H800 vs NVIDIA RTX 5090 for local LLMs
0 reports on the NVIDIA H800 and 168 on the NVIDIA RTX 5090
Which is faster for local LLMs?
Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 1.9x higher on the NVIDIA H800 (3,350 GB/s on the NVIDIA H800, 1,792 on the NVIDIA RTX 5090). The estimates below compare them on the models people run most.
NVIDIA H800
- Memory
- 80GB
- Memory bandwidth
- 3,350 GB/s
- FP16 compute
- 989.5 TFLOPS
- Reports
- 0
NVIDIA RTX 5090
- Memory
- 32GB
- Memory bandwidth
- 1,792 GB/s
- FP16 compute
- 419 TFLOPS
- Reports
- 168
Estimated by the calculator
The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape.
| Model | NVIDIA H800 | NVIDIA RTX 5090 |
|---|---|---|
| Qwen3.8 27B | 34.1 t/sFP16 | 44.9 t/sQ6_K |
| Qwen3.8 125B · 6B active | 202.0 t/sQ4_K_S | 20.3 t/sQ2_K, experts in RAM |
| Qwen3.6 35B · 3B active | 245.7 t/sQ8_0 | 163.0 t/sQ5_K_M |
| DeepSeek V4 Flash 284B · 13B active | 9.6 t/sQ2_K, experts in RAM | doesn't fit at these settings |
| Qwen3.6 27B | 34.1 t/sFP16 | 44.9 t/sQ6_K |
| Gemma 4 26B · 4B active | 87.8 t/sFP16 | 114.5 t/sQ6_K |
Change the context, RAM or model in the calculator: NVIDIA H800 · NVIDIA RTX 5090
Every report on each card, with its source: NVIDIA H800 · NVIDIA RTX 5090