community pool / model answer

Qwen3.8-27B-Uncensored-FP8

chat · orcarouter/Qwen3.8-27B-Uncensored-FP8

128.4 tok/s · model median · 5 runs in the model record.

Answer

113.7tok/s

measuredRTX PRO 6000 Blackwell 96GB8-bit · 3 runs · decodebest cell in the pool

  1. RTX PRO 6000 Blackwell Workstation Edition 96GB166.6 tok/sreported2 runs · peak 76.9 GB · fits 96 GB rig memory
  2. RTX PRO 6000 Blackwell 96GB113.7 tok/smeasured3 runs · peak VRAM no data yet · 96 GB rig memory
Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX PRO 6000 Blackwell 96GB8-bit113.7 tok/smeasured108- GB524,288sglang
RTX PRO 6000 Blackwell Workstation Edition 96GB8-bit166.6 tok/sreported7176.9 GB2,048sglang

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.