2× NVIDIA RTX A5000 24 GB
3 records across 1 models (quants: FP8).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2× NVIDIA RTX A5000 24 GB | Qwen3-35B-A3B | FP8 | vLLM | 145.3 | – | – | – | – | 2026-09-27 | community | – | source |
| 2× NVIDIA RTX A5000 24 GB | Qwen3-35B-A3B | FP8 | vLLM | 48.6 | – | – | – | 163840 | 2026-09-27 | community | – | source |
| 2× NVIDIA RTX A5000 24 GB | Qwen3-35B-A3B | FP8 | vLLM | 11 | – | – | – | – | 2026-09-27 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085