bestmodel.run / model — Qwen3.6-35B-A3B (agent view) category chat · Qwen/Qwen3.6-35B-A3B · 137 runs in the model record answer: 71.3 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 Arc Pro B60 24GB | 4-bit | 42.6 tok/s | reported | ttft 590ms | vram - GB | ctx 2,048 Arc Pro B70 32GB | 4-bit | 70 tok/s | measured | ttft 372ms | vram 22 GB | ctx 32,768 Arc Pro B70 32GB | 5-bit | 67 tok/s | measured | ttft 69ms | vram - GB | ctx 262,144 Arc Pro B70 32GB ×2 | 4-bit | 73.9 tok/s | reported | ttft -ms | vram - GB | ctx 2,048 Arc Pro B70 32GB ×2 | 8-bit | 110.7 tok/s | reported | ttft -ms | vram 31.2 GB | ctx 65,536 Arc Pro B70 32GB ×4 | 1-bit | 120.8 tok/s | reported | ttft -ms | vram 29.8 GB | ctx 65,536 Arc Pro B70 32GB ×4 | 4-bit | 68.8 tok/s | reported | ttft -ms | vram 60 GB | ctx 262,144 Arc Pro B70 32GB ×4 | 8-bit | 99.8 tok/s | reported | ttft 77ms | vram 127.6 GB | ctx 32,768 Arc Pro B70 32GB ×4 | 16-bit | 102.5 tok/s | reported | ttft 104ms | vram 123 GB | ctx 262,144 Intel(R) Core(TM) Ultra 7 155H | 3-bit | 13.4 tok/s | reported | ttft -ms | vram - GB | ctx 2,048 GB10 Grace Blackwell 128GB | 4-bit | 64.5 tok/s | reported | ttft 153ms | vram 87 GB | ctx 262,144 GB10 Grace Blackwell 128GB | 6-bit | 56.1 tok/s | reported | ttft -ms | vram - GB | ctx 2,048 GB10 Grace Blackwell 128GB | 8-bit | 55.7 tok/s | measured | ttft 206ms | vram - GB | ctx 262,144 GTX 1060 6GB | 2-bit | 15 tok/s | reported | ttft -ms | vram 5.9 GB | ctx 76,800 GTX 1080 Ti 11GB | 2-bit | 21.6 tok/s | measured | ttft -ms | vram 10.4 GB | ctx 8,192 GTX 1080 Ti 11GB | 3-bit | 14.6 tok/s | reported | ttft -ms | vram 9.6 GB | ctx 8,192 GTX 1080 Ti 11GB | 4-bit | 10.3 tok/s | reported | ttft -ms | vram 10.2 GB | ctx 8,192 GTX 1080 Ti 11GB | 8-bit | 6.4 tok/s | reported | ttft -ms | vram 10.9 GB | ctx 8,192 M1 Max 64GB | 4-bit | 59.3 tok/s | reported | ttft 45ms | vram - GB | ctx 2,048 M3 Ultra 512GB | 4-bit | 31.9 tok/s | reported | ttft 620ms | vram - GB | ctx 2,048 M5 Max 128GB | 4-bit | 105.2 tok/s | measured | ttft 28ms | vram 19.9 GB | ctx 8,192 Multi-GPU 28GB ×2 | 4-bit | 85.6 tok/s | measured | ttft 81ms | vram - GB | ctx 4,096 Radeon AI Pro R9700 32GB | 4-bit | 277.4 tok/s | measured | ttft 520ms | vram - GB | ctx 4,096 Radeon AI Pro R9700 32GB ×2 | 4-bit | 32.6 tok/s | reported | ttft 135ms | vram - GB | ctx 32,768 Radeon AI Pro R9700 32GB ×3 | 4-bit | 98.9 tok/s | measured | ttft 171ms | vram 32 GB | ctx 787 Radeon AI Pro R9700 32GB ×3 | 8-bit | 79.2 tok/s | measured | ttft 343ms | vram 44.8 GB | ctx 787 Radeon AI Pro R9700 34GB | 4-bit | 259.5 tok/s | reported | ttft 77ms | vram 27.8 GB | ctx 4,096 Radeon AI Pro R9700 57GB | 4-bit | 33.2 tok/s | reported | ttft 1,294ms | vram - GB | ctx 2,048 Radeon AI Pro R9700 57GB | 6-bit | 26.4 tok/s | reported | ttft 1,568ms | vram - GB | ctx 2,048 Radeon AI Pro R9700 57GB | 8-bit | 24.4 tok/s | reported | ttft 1,254ms | vram - GB | ctx 2,048 Radeon RX 9070 16GB | 4-bit | 34.2 tok/s | reported | ttft 417ms | vram 14.6 GB | ctx 131,072 RTX 3050 8GB | 3-bit | 31 tok/s | reported | ttft 139ms | vram - GB | ctx 28,000 RTX 3060 12GB | 3-bit | 27.9 tok/s | measured | ttft 1,476ms | vram 5.6 GB | ctx 4,096 RTX 3060 12GB | 4-bit | 35 tok/s | reported | ttft -ms | vram - GB | ctx 131,072 RTX 3060 12GB ×2 | 4-bit | 67.6 tok/s | reported | ttft -ms | vram 20.9 GB | ctx 96,000 RTX 3070 Ti 8GB | 3-bit | 33.5 tok/s | reported | ttft -ms | vram 7,949 GB | ctx 262,144 RTX 3080 12GB | 4-bit | 35.4 tok/s | measured | ttft -ms | vram 9.5 GB | ctx 196,608 RTX 3090 24GB | 4-bit | 11 tok/s | reported | ttft 647ms | vram - GB | ctx 2,048 RTX 3090 24GB ×2 | 4-bit | 155.6 tok/s | measured | ttft 121ms | vram 21.3 GB | ctx 32,768 RTX 3090 24GB ×2 | 6-bit | 137.9 tok/s | reported | ttft 648ms | vram - GB | ctx 2,048 RTX 4070 12GB | 4-bit | 39.6 tok/s | measured | ttft 1,476ms | vram - GB | ctx 2,048 RTX 4090 24GB | 3-bit | 157.7 tok/s | reported | ttft 175ms | vram - GB | ctx 8,192 RTX 5080 16GB | 3-bit | 150.6 tok/s | reported | ttft -ms | vram 14.2 GB | ctx 131,072 RTX 5090 32GB | 4-bit | 190 tok/s | measured | ttft 54ms | vram 26.7 GB | ctx 262,144 RTX 5090 32GB | 6-bit | 219.8 tok/s | reported | ttft 21ms | vram 27.5 GB | ctx 4,096 RTX PRO 6000 Blackwell Workstation Edition 96GB | 16-bit | 164.1 tok/s | reported | ttft 3,449ms | vram 85.3 GB | ctx 65,536 RX 6700 10GB | 3-bit | 16.2 tok/s | reported | ttft 2,448ms | vram 4 GB | ctx 4,096 RX 7900 XTX 24GB | 4-bit | 175 tok/s | measured | ttft 76ms | vram 22.1 GB | ctx 4,096 Ryzen AI Max 395 103GB | 4-bit | 174.9 tok/s | reported | ttft 296ms | vram - GB | ctx 4,096 Ryzen AI Max 395 64GB | 4-bit | 62.4 tok/s | measured | ttft 1,020ms | vram 21.5 GB | ctx 8,192 T4 16GB | 2-bit | 54.7 tok/s | reported | ttft -ms | vram - GB | ctx 2,048 Tesla P100 32GB ×2 | 4-bit | 78.1 tok/s | reported | ttft 157ms | vram - GB | ctx 2,048 Tesla V100 32GB ×2 | 6-bit | 96.7 tok/s | reported | ttft -ms | vram 60 GB | ctx 65,536 Tesla V100 SXM2 32GB | 4-bit | 66.8 tok/s | reported | ttft 40,000ms | vram 28 GB | ctx 96,000 Ultra 7 155H 32GB | 3-bit | 8.7 tok/s | reported | ttft -ms | vram - GB | ctx 2,048