community pool / model answer

Qwen3.8-27B-NVFP4

chat · RadixArk/Qwen3.8-27B-NVFP4

35.3 tok/s · model median · 4 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
GB10 Grace Blackwell 128GB4-bit23.6 tok/smeasured226- GB131,072sglang, vllm
RTX PRO 6000 Blackwell 96GB4-bit97.4 tok/sreported40- 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.