M5 Pro (20-core GPU)
16 records across 1 models (quants: Q8_0).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | 9.25 | – | – | – | 8192 | 2026-10 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | 9.18 | – | – | – | – | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | 9.09 | – | – | – | 8192 | 2026-10 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | 9.02 | – | – | – | – | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 383.3 | – | – | – | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 385.8 | – | – | – | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 295.7 | – | – | 8192 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 340.0 | – | – | 8192 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 174.3 | – | – | 32768 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 252.9 | – | – | 32768 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 107.2 | – | – | 8192 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp | – | 106.2 | – | – | 8192 | 2026-09-28 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | – | 323.8 | – | – | 8192 | 2026-10 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | – | 355.3 | – | – | 8192 | 2026-10 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | – | 184.2 | – | – | 32768 | 2026-10 | community | – | source |
| M5 Pro (20-core GPU) | Qwen3.8-27B | Q8_0 | llama.cpp (Metal) | – | 265.6 | – | – | 32768 | 2026-10 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085