community pool / model answer
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4
chat · nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4
494.7 tok/s · model median · 8 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| RTX 3090 24GB ×2 | 4-bit | 197 tok/s | reported | 8,037 | 19.5 GB | 131,072 | vllm |
| RTX 5090 32GB | 4-bit | 564.1 tok/s | reported | 76 | 30.1 GB | 65,536 | vllm |
| RTX PRO 6000 Blackwell 96GB | 4-bit | 494.7 tok/s | measured | 34 | 86.1 GB | 32,768 | vllm |
Context tested
The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.