M1 Max (64 GB)
10 records across 5 models (quants: 4bit, IQ4_XS, Q4_K - Medium, Q4_K_XL, Q6_K, Q8_0).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M1 Max (64 GB) | GLM-4.7-Flash | Q4_K_XL | llama.cpp | 36.8 | 99.4 | – | – | 4096 | – | sourced | – | source |
| M1 Max (64 GB) | Nemotron 3 Nano 30B-A3B | Q4_K_XL | llama.cpp | 43.7 | 136.9 | – | – | 4096 | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3-Coder-30B-A3B | IQ4_XS | llama.cpp | 58.5 | 132.1 | – | – | 4096 | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-27B | 4bit | MLX | 15 | 67 | – | – | 8000 | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-27B | Q6_K | llama.cpp | 12 | – | – | – | – | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-27B | Q4_K - Medium | llama.cpp | 11.5 | – | – | – | – | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-27B | Q8_0 | llama.cpp | 10.5 | – | – | – | – | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-35B-A3B | 4bit | MLX | 57.6 | 431 | – | – | 8000 | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-35B-A3B | 4bit | LM Studio (MLX) | 57 | – | – | – | – | – | sourced | – | source |
| M1 Max (64 GB) | Qwen3.5-35B-A3B | Q4_K - Medium | LM Studio (llama.cpp) | 29 | – | – | – | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085