llama 8B
6 records across 2 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Arc(TM) B390 | llama 8B 8.03B | Q4_0 | llama.cpp (Vulkan) | 24.45 | 1380.07 | – | – | tg128 | 2026-09-12 | community | – | source |
| Arc(TM) B390 | llama 8B 8.03B | Q4_0 | llama.cpp (Vulkan) | 23.93 | 1332.41 | – | – | tg128 | 2026-09-12 | community | – | source |
| Arc(TM) B390 | llama 8B 8.03B | Q4_0 | llama.cpp (Vulkan) | 23.47 | 733.92 | – | – | tg2048 | 2026-09-12 | community | – | source |
| Arc(TM) B390 | llama 8B 8.03B | Q4_0 | llama.cpp (Vulkan) | 22.04 | 314.3 | – | – | tg2048 | 2026-09-12 | community | – | source |
| RTX 3060 | llama 8B | Q4_K - Medium | llama.cpp (Vulkan) | 12.11 | 230.26 | – | – | tg128 | 2025-11-18 | community | – | source |
| RTX 3060 | llama 8B | Q4_K - Medium | llama.cpp (Vulkan) | 9.06 | 1061.32 | – | – | tg128 | 2025-11-18 | community | – | source |