bestmodel.run / model — Qwen3.8-27B-GGUF (agent view) category chat · unsloth/Qwen3.8-27B-GGUF · 519 runs in the model record answer: 44 tok/s (model median) Honesty ladder: measured > reported > extrapolated > formula > no data yet. Empty fields stay empty — never filled with an estimate. rig | quant | median result | basis | ttft | peak vram | max context A100 40GB | 4-bit | 41.1 tok/s | reported | ttft 10,667ms | vram 19 GB | ctx 65,536 A100 40GB | 8-bit | 33.7 tok/s | reported | ttft 10,194ms | vram 30 GB | ctx 65,536 Arc Pro B60 24GB | 4-bit | 21.2 tok/s | reported | ttft 147ms | vram - GB | ctx 2,048 Arc Pro B60 24GB | 6-bit | 16 tok/s | reported | ttft 796ms | vram - GB | ctx 2,048 Arc Pro B70 32GB | 4-bit | 31.2 tok/s | reported | ttft 2,995ms | vram - GB | ctx 2,048 Arc Pro B70 32GB | 6-bit | 24.1 tok/s | reported | ttft 354ms | vram - GB | ctx 2,048 Arc Pro B70 32GB | 8-bit | 16.1 tok/s | reported | ttft 572ms | vram - GB | ctx 2,048 B200 192GB | 4-bit | 86.8 tok/s | reported | ttft 632ms | vram 17.3 GB | ctx 32,768 CMP 170HX 64GB | 4-bit | 47 tok/s | measured | ttft 29,673ms | vram - GB | ctx 132,352 CMP 170HX 64GB | 8-bit | 49.7 tok/s | measured | ttft 27,843ms | vram - GB | ctx 132,352 CMP 170HX 64GB | 16-bit | 23.2 tok/s | measured | ttft 13,004ms | vram - GB | ctx 67,549 M4 Pro 24GB | 4-bit | 47.2 tok/s | reported | ttft 373ms | vram - GB | ctx 2,048 M5 Max 128GB | 4-bit | 24.3 tok/s | reported | ttft 85ms | vram - GB | ctx 2,048 Multi-GPU 56GB ×2 | 8-bit | 36.7 tok/s | reported | ttft 3,082ms | vram 45.5 GB | ctx 2,048 Radeon AI Pro R9700 32GB | 1-bit | 41.2 tok/s | measured | ttft 455ms | vram 7.3 GB | ctx 32,768 Radeon AI Pro R9700 32GB | 4-bit | 39.5 tok/s | measured | ttft 615ms | vram 21.6 GB | ctx 262,144 Radeon AI Pro R9700 32GB | 6-bit | 21.6 tok/s | measured | ttft 641ms | vram 25.6 GB | ctx 32,768 RTX 3060 12GB | 2-bit | 18.3 tok/s | measured | ttft 150ms | vram - GB | ctx 32,768 RTX 3060 12GB ×2 | 3-bit | 15.5 tok/s | reported | ttft -ms | vram - GB | ctx 32,768 RTX 3060 12GB ×2 | 4-bit | 18.7 tok/s | measured | ttft 1,506ms | vram 19.4 GB | ctx 131,072 RTX 3080 12GB | 2-bit | 44.4 tok/s | measured | ttft 2,526ms | vram 9.7 GB | ctx 65,536 RTX 3090 24GB | 4-bit | 72.6 tok/s | measured | ttft 647ms | vram 18.4 GB | ctx 32,768 RTX 3090 24GB | 6-bit | 51.5 tok/s | reported | ttft 206ms | vram - GB | ctx 2,048 RTX 3090 24GB ×2 | 4-bit | 73 tok/s | reported | ttft 1,175ms | vram - GB | ctx 2,048 RTX 3090 24GB ×2 | 8-bit | 51.6 tok/s | measured | ttft 230ms | vram - GB | ctx 262,144 RTX 4060 Ti 16GB ×2 | 4-bit | 53.6 tok/s | reported | ttft -ms | vram - GB | ctx 2,048 RTX 4070 12GB | 2-bit | 65.7 tok/s | reported | ttft 227ms | vram 9.3 GB | ctx 2,048 RTX 5070 12GB | 2-bit | 46.3 tok/s | measured | ttft 572ms | vram 9.8 GB | ctx 32,768 RTX 5090 32GB | 2-bit | 88.1 tok/s | reported | ttft 82ms | vram - GB | ctx 2,048 RTX 5090 32GB | 3-bit | 79.4 tok/s | measured | ttft 382ms | vram 17.3 GB | ctx 131,072 RTX 5090 32GB | 4-bit | 101.6 tok/s | measured | ttft 130ms | vram 21.3 GB | ctx 262,144 RTX 5090 32GB | 5-bit | 60.2 tok/s | measured | ttft 424ms | vram 23.2 GB | ctx 131,072 RTX 5090 32GB | 6-bit | 79.3 tok/s | measured | ttft 423ms | vram 30.3 GB | ctx 262,144 RTX PRO 6000 Blackwell 96GB | 2-bit | 100.7 tok/s | measured | ttft -ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 3-bit | 88.9 tok/s | measured | ttft -ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 4-bit | 79.3 tok/s | measured | ttft -ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 5-bit | 67.8 tok/s | measured | ttft -ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 6-bit | 57.5 tok/s | reported | ttft -ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 8-bit | 50.2 tok/s | measured | ttft 42,652ms | vram - GB | ctx 131,072 RTX PRO 6000 Blackwell 96GB | 16-bit | 61 tok/s | measured | ttft 40,786ms | vram - GB | ctx 131,072 RX 7800 XT 16GB | 4-bit | 42 tok/s | reported | ttft -ms | vram - GB | ctx 121,000 RX 7900 XTX 24GB | 4-bit | 42.3 tok/s | measured | ttft 298ms | vram 22.9 GB | ctx 180,224 Ryzen AI Max 395 128GB | 4-bit | 26.7 tok/s | measured | ttft 603ms | vram 23.9 GB | ctx 131,072 Ryzen AI Max 395 128GB | 5-bit | 13.4 tok/s | reported | ttft 642ms | vram 23.1 GB | ctx 4,096 Ryzen AI Max 395 128GB | 6-bit | 13.3 tok/s | reported | ttft 730ms | vram 25.7 GB | ctx 4,096 Ryzen AI Max 395 128GB | 8-bit | 11.2 tok/s | reported | ttft 746ms | vram 31.1 GB | ctx 4,096 Ryzen AI Max 395 64GB | 4-bit | 24.1 tok/s | measured | ttft 4,332ms | vram 18 GB | ctx 8,192