Qwen3.5-0.8B-Japanese-SFT
4 records across 1 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 | Qwen3.5-0.8B-Japanese-SFT | Q4_K_M | llama.cpp (Metal) | 169 | – | – | – | tg128 | 2026-10-02 | community | – | source |
| M5 | Qwen3.5-0.8B-Japanese-SFT | Q8_0 | llama.cpp (Metal) | 123 | – | – | – | tg128 | 2026-10-02 | community | – | source |
| M5 | Qwen3.5-0.8B-Japanese-SFT | Q4_K_M | llama.cpp (CPU) | 92 | – | – | – | tg128 | 2026-10-02 | community | – | source |
| M5 | Qwen3.5-0.8B-Japanese-SFT | Q8_0 | llama.cpp (CPU) | 78 | – | – | – | tg128 | 2026-10-02 | community | – | source |