GPU comparison
MacBook Air M4 (24GB) vs MacBook Air/Pro M5 (16GB) for local LLMs
How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 64 tracked models each one can run.
| MacBook Air M4 (24GB) | MacBook Air/Pro M5 (16GB) | |
|---|---|---|
| Memory | 24.0 GB | 16.0 GB |
| Bandwidth | 120 GB/s | 153 GB/s |
| Price (approx) | $1,399 | $1,099 |
| LLMs it runs | 45 of 64 | 27 of 64 |
| Best model it runs | Qwen3.8 27B · 7–10 tok/s | Devstral Small 2 24B · 10–15 tok/s |
The MacBook Air M4 (24GB) runs 18 more of the tracked models (45 vs 27), thanks to its 24.0 GB of memory. The MacBook Air/Pro M5 (16GB) has more memory bandwidth (153 GB/s), so it generates tokens faster at the same model and quant.