community pool / model answer

Qwen3.8-27B-exl3

chat · turboderp/Qwen3.8-27B-exl3

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

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX 5070 12GB3-bit71.9 tok/smeasured2,36410 GB126,976exllamav3

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.