4x GB200
12 records across 2 models (quants: NVFP4).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 4x GB200 | GLM-5.2-NVFP4 | NVFP4 | vLLM | 340.987 | – | – | – | – | 2026-10-08 | community | – | source |
| 4x GB200 | GLM-5.2-NVFP4 | NVFP4 | vLLM | 340.973 | – | – | – | – | 2026-10-08 | community | – | source |
| 4x GB200 | GLM-5.2-NVFP4 | NVFP4 | vLLM | 47.538 | – | – | – | – | 2026-10-08 | community | – | source |
| 4x GB200 | GLM-5.2-NVFP4 | NVFP4 | vLLM | 46.576 | – | – | – | – | 2026-10-08 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 10371 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 9109 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 7174 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 5476 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 2178 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 1879 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 317 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
| 4x GB200 | Nemotron-3.5-Lightning 30B-A3B | NVFP4 | vLLM | 277 | – | – | – | tg1024 | 2026-10-05 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085