NVIDIA H100 NVL vs NVIDIA Jetson AGX Thor 128GB for local LLMs
0 reports on the NVIDIA H100 NVL and 5 on the NVIDIA Jetson AGX Thor 128GB
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 14.3x higher on the NVIDIA H100 NVL (3,900 GB/s on the NVIDIA H100 NVL, 273 on the NVIDIA Jetson AGX Thor 128GB). The estimates below compare them on the models people run most.
- Memory
- 94GB
- Memory bandwidth
- 3,900 GB/s
- FP16 compute
- 835.5 TFLOPS
- Reports
- 0
- Memory
- 128GB unified
- Memory bandwidth
- 273 GB/s
- FP16 compute
- 258.8 TFLOPS
- Reports
- 5
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 H100 NVL | NVIDIA Jetson AGX Thor 128GB |
|---|---|---|
| Qwen3.8 27B | 39.7 t/sFP16 | 2.8 t/sFP16 |
| Qwen3.8 125B · 6B active | 217.3 t/sQ4_K_M | 10.0 t/sQ6_K |
| DeepSeek V4 Flash 284B · 13B active | 9.6 t/sQ2_K, experts in RAM | 26.5 t/sQ2_K |
| Qwen3.6 35B · 3B active | 123.0 t/sFP16 | 10.0 t/sFP16 |
| Qwen3.6 27B | 39.7 t/sFP16 | 2.8 t/sFP16 |
| Gemma 4 26B · 4B active | 107.3 t/sFP16 | 7.5 t/sFP16 |
Change the context, RAM or model in the calculator: NVIDIA H100 NVL · NVIDIA Jetson AGX Thor 128GB
Every report on each card, with its source: NVIDIA H100 NVL · NVIDIA Jetson AGX Thor 128GB