bestmodel.run / model — DeepSeek-V4-Flash (agent view) category chat · deepseek-ai/DeepSeek-V4-Flash · 36 runs in the model record answer: 28.5 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 GB10 Grace Blackwell 128GB | 8-bit | 45.7 tok/s | reported | ttft 152ms | vram 209 GB | ctx 200,000 GB10 Grace Blackwell 256GB | 4-bit | 26.9 tok/s | reported | ttft 495ms | vram - GB | ctx 393,216 GB10 Grace Blackwell 256GB | 8-bit | 36 tok/s | reported | ttft 1,202ms | vram 99.2 GB | ctx 500,000 M3 Ultra 512GB | 2-bit | 33 tok/s | reported | ttft 300ms | vram - GB | ctx 16,384 M5 Max 128GB | 2-bit | 19 tok/s | measured | ttft 70,150ms | vram - GB | ctx 128,000 Radeon 8060S 103.1GB | 2-bit | 19 tok/s | reported | ttft 659ms | vram 83 GB | ctx 4,096 Radeon AI Pro R9700 32GB ×3 | 1-bit | 9.2 tok/s | measured | ttft 4,403ms | vram 81.6 GB | ctx 8,192 Radeon AI Pro R9700 32GB ×4 | 2-bit | 53.7 tok/s | reported | ttft 5,602ms | vram - GB | ctx 2,052 RTX 3090 24GB ×4 | 2-bit | 35.3 tok/s | reported | ttft 1,283ms | vram 83.9 GB | ctx 2,048 RTX 5090 32GB ×4 | 2-bit | 212 tok/s | reported | ttft 55ms | vram 124 GB | ctx 262,144 RTX PRO 6000 Blackwell 96GB | 3-bit | 80.8 tok/s | reported | ttft 9,240ms | vram 94.5 GB | ctx 262,144 RTX PRO 6000 Blackwell 96GB ×2 | 4-bit | 160.9 tok/s | reported | ttft 1,130ms | vram 188.8 GB | ctx 500,000 RTX PRO 6000 Blackwell 96GB ×2 | 8-bit | 261.8 tok/s | reported | ttft 4,830ms | vram 184 GB | ctx 262,000 Ryzen AI Max 395 103GB | 2-bit | 34.9 tok/s | reported | ttft 1,159ms | vram - GB | ctx 4,096 Ryzen AI Max 395 128GB | 2-bit | 15.2 tok/s | measured | ttft -ms | vram - GB | ctx 65,536