community pool / model answer

Laguna-S-2.1-DFlash-FP8

chat · poolside/Laguna-S-2.1-DFlash-FP8

196.3 tok/s · model median · 2 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX PRO 6000 Blackwell 96GB ×28-bit182.7 tok/sreported58- GB2,048vllm
RTX PRO 6000 Blackwell 96GB ×48-bit209.9 tok/sreported49- GB2,048vllm

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.