The best local LLM for a MacBook Air M3 (24 GB) is Qwen3.8-27B at 5 tok/s (estimated). With 24 GB of unified memory it runs 43 of the models in the LLMCheck index — from compact options up to 35B-class models. For everyday chat and coding, Qwen3.8-27B is the sweet spot. Full ranking below.
Considering this Mac?
Mac mini M5 Pro on Amazon → · compare every Mac for local AI →
Ranked by LLMCheck suitability (capability balanced against speed on the M3). Click a model for its full benchmark and setup. Every speed carries its provenance mark: most are index estimates from the memory-bandwidth model, and a filled square is a published measurement — submit a real run →
| # | Model | Size | License | Speed | Capability |
|---|---|---|---|---|---|
| 1 | Qwen3.8-27B | 27.8B | Apache 2.0 | estimated5 tok/s | 46/50 |
| 2 | Qwen 3.6-27B | 27B | Apache 2.0 | estimated5 tok/s | 44/50 |
| 3 | Muse Glimmer 30B | 30B | Apache 2.0 | estimated5 tok/s | 42/50 |
| 4 | Gemma 4 31B | 31B | Apache 2.0 | estimated4 tok/s | 40/50 |
| 5 | KAT-Coder-V2.5 | 35B | Apache 2.0 | estimated9 tok/s | 39/50 |
| 6 | Qwen 3.6-35B-A3B | 35B | Apache 2.0 | estimated8 tok/s | 38/50 |
| 7 | Nemotron 3.5 Lightning | 30B | OpenMDW | estimated10 tok/s | 36/50 |
| 8 | Gemma 4 26B-A4B | 26B | Apache 2.0 | estimated8 tok/s | 35/50 |
| 9 | GLM-4.7-Flash | 31B | MIT | estimated4 tok/s | 35/50 |
| 10 | Maple Preview 20B-A1B | 20B | MIT | estimated175 tok/s | 20/50 |
| 11 | Laguna XS 2.1 | 33B | OpenMDW | estimated9 tok/s | 33/50 |
| 12 | Bonsai 27B | 27B | Apache 2.0 | estimated5 tok/s | 33/50 |
Showing the top 12 of 43 models that fit in 24 GB. See the full leaderboard or all benchmarks.
The fastest way to get started is Ollama. Install it, then find the top pick for your Mac in its library and copy the exact tag it lists:
Find Qwen3.8-27B in the Ollama library → · model card on Hugging Face →
Prefer a GUI? LM Studio gives you a one-click download and chat window. For step-by-step help see our Ollama install guide, or open the Qwen3.8-27B on M3 benchmark page for exact settings.
The MacBook Air M3 (24 GB) tops out at KAT-Coder-V2.5. Newer Apple Silicon with more unified memory runs larger, smarter models much faster:
As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence our rankings.
Qwen3.8-27B (27.8B, Apache 2.0) is the best all-round pick at 5 tok/s (estimated) on the M3. If you want maximum speed, Maple Preview 20B-A1B hits 175 tok/s (estimated); for maximum capability, Qwen 3.6-27B still fits in 24 GB.
About 43 of the 89 models in the LLMCheck leaderboard fit in 24 GB of unified memory, from compact models up to KAT-Coder-V2.5 (35B).
Not comfortably. A 70B model in Q4 needs ~40–44 GB; with 24 GB you should stick to models up to ~18 GB, such as Qwen3.8-27B. For 70B, look at a 48 GB+ Mac.
24 GB is great for small-to-mid models (up to ~14B comfortably); for 30B+ you'll want 32 GB or more. Because Apple Silicon uses unified memory, that figure is both your system RAM and your VRAM.