The best local LLM for a 16 GB Mac is Gemma 4 12B, a 7.7 GB download on a Mac that reads images and fits with room to spare. The smaller Gemma 4 E4B is a tight fit, because Ollama gives Macs a 9.5 GB version of it, but it answered at 87 tokens a second in my tests. For coding agents, Ornith 1.0 9B (5.6 GB) is built for small machines. All of these run on the Mac itself, with nothing sent to a vendor, in free apps such as Ollama and LM Studio.
Key facts
- Install first: Gemma 4 12B, 7.7 GB download.
- Room for a model: about 12 GB of 16 GB, so a file under about 8 GB runs comfortably.
- App: Ollama or LM Studio, both free. The commands below are for Ollama.
- Cost: nothing. Check each licence before business use.
On this page
- Models that fit a 16 GB Mac
- Which one for which job
- How much room a 16 GB Mac gives a model
- Too big for 16 GB
- Frequently asked questions
Models that fit a 16 GB Mac
| Model | Good for | Download | Fit on 16 GB | Licence | Command |
|---|---|---|---|---|---|
| Gemma 4 12B | Chat, coding, reads images | 7.7 GB | Comfortable | Apache 2.0 | ollama run gemma4:12b |
| Gemma 4 E4B | Everyday chat, reads images, calls tools | 9.5 GB | Tight | Apache 2.0 | ollama run gemma4 |
| Qwen3.5 9B | Coding, reasoning, reads images | 7.6 GB | Comfortable | Apache 2.0 | ollama run qwen3.5:9b |
| Ornith 1.0 9B | Coding agents on small machines | 5.6 GB | Comfortable | MIT | ollama run ornith:9b |
| Qwen3 8B | Chat, coding | 5.2 GB | Comfortable | Apache 2.0 | ollama run qwen3:8b |
| Ministral 3 14B | Chat, reasoning, reads images | 9.1 GB | Tight | Apache 2.0 | ollama run ministral-3:14b |
Comfortable means the file fits with room for a long conversation. Tight means it loads, but close other apps and keep chats short. Sizes are what Ollama downloads on a Mac, read from its registry on 21 November 2026. Where Apple's MLX build of a model exists, Ollama now uses it on a Mac, so some sizes differ from the ones on Ollama's website. On a Mac, Gemma 4 12B is a smaller download than Gemma 4 E4B, because Ollama sends each a different build.
Tested: on my 24 GB M5 MacBook Air, Gemma 4 E4B answered at 87 tokens a second and used 9.4 GB of memory. A 16 GB Mac with the same chip should be close to that speed, as long as you close other apps.
Which one for which job
- Everyday chat, summaries and images: Gemma 4 12B.
- Fast answers, with other apps closed: Gemma 4 E4B.
- Coding agents: Ornith 1.0 9B.
- Coding and reasoning: Qwen3.5 9B.
To check any other model against your Mac, use the local LLM memory calculator. For every Mac size on one page, see the best local LLM for Mac.
How much room a 16 GB Mac gives a model
A Mac shares its memory between the processor and the graphics side, and by default macOS lets the graphics side use about 75% of it. On a 16 GB Mac that is about 12 GB. The conversation needs 2 to 4 GB on top of the model file, so a file under about 8 GB runs comfortably and one up to about 10 GB runs if you keep chats short.
Longer documents and longer chats use more memory, so leave more room if you paste in long reports. Speed depends mostly on the chip: a Pro or Max chip moves data faster than a base chip, so the same model answers faster.
Too big for 16 GB
gpt-oss 20B (13.8 GB): this needs more room than a 16 GB Mac gives. See the best local LLM for a 24 GB Mac.
Which app to run it in
Ollama runs models in the background for other apps and agents to use; the commands above install and start each model. LM Studio gives you a chat window and a model search. Both are free. See Ollama vs LM Studio if you can't decide.
Frequently asked questions
What is the best local LLM for a 16 GB Mac?
The best local LLM for a 16 GB Mac is Gemma 4 12B, a 7.7 GB download on a Mac that reads images and fits with room to spare. The smaller Gemma 4 E4B is a tight fit, because Ollama gives Macs a 9.5 GB version of it, but it answered at 87 tokens a second in my tests. For coding agents, Ornith 1.0 9B (5.6 GB) is built for small machines.
How big a model can a 16 GB Mac run?
By default macOS lets a model use about 12 GB of a 16 GB Mac. Allow 2 to 4 GB for the conversation, so aim for a model file under about 8 GB, or up to 10 GB with short chats.
Can a 16 GB Mac run gpt-oss 20B?
Not comfortably. gpt-oss 20B is a 13.8 GB download, more than the roughly 10 GB that leaves room for a conversation. It needs a 24 GB Mac.
Are these models free?
Yes, to download and run, with nothing sent to a vendor. Check the licence before business use: Apache 2.0 and MIT allow commercial use, while the Llama, LFM and NVIDIA licences add their own conditions.
Sources
- Ollama library and registry: model pages, and the download size for a Mac (read 21 November 2026). On my M5 MacBook Air, Ollama 0.40.17 chose the MLX build for Gemma 4 E4B.
- ModelFit's open hardware dataset (CC BY 4.0), updated 26 October 2026. Good-for labels adapted; fit worked out with my own memory rule.
- Local LLM memory calculator, for the 75% memory rule
Other Mac sizes: 8 GB · 24 GB · 32 GB · 48 GB · 64 GB · 128 GB.
New models arrive every week. See Local AI This Week for the latest. Spotted something out of date? Tell me.



