M4 Max 96GB vs M5 Max 128GB
For running local LLMs · 18 reports across 9 models
Highest reported generation speed by model family
These maxima come from different community setups. Model size, quantization, GPU count, offloading and concurrency may differ. They do not establish which GPU is faster for the same workload. Check the comparison conditions →
| Model | M4 Max 96GB | M5 Max 128GB |
|---|---|---|
| Qwen3.8up to 2400B | — | 55.0n=5 |
| Qwen3.6up to 35B | — | 79.4n=4 |
| Gemma 4up to 31B | — | 7.5n=3 |
| DeepSeek V4 Flashup to 285B | — | —n=1 |
| GLM-5.3up to 321B | — | 24.0n=1 |
| Nex-N2.5-mini | — | 133.6n=1 |
| Qwen3.5up to 122B | — | —n=1 |
| StepFun 3.7 | — | 33.9n=1 |
| Tencent-HY3up to 295B | — | 32.4n=1 |