community pool / model answer
Qwen3.8-27B-W4A16-AutoRound
chat · dbirks/Qwen3.8-27B-W4A16-AutoRound
187.5 tok/s · model median · 14 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| CMP 170HX 64GB | 4-bit | 255.9 tok/s | measured | 440 | 56.5 GB | 131,072 | vllm |
| RTX 3090 24GB ×2 | 4-bit | 182.4 tok/s | measured | 825 | 44.4 GB | 262,144 | sglang, vllm |
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.