Memory at each precision
| Version | Weights | With context | Smallest Mac |
|---|---|---|---|
| 8-bit | 126 GB | 137 GB | 192 GB |
| 6-bit | 96.7 GB | 104 GB | 192 GB |
| 4-bit | 66.9 GB | 72.3 GB | 128 GB |
| 3-bit lower quality | 52.1 GB | 56.3 GB | 96 GB |
| 2-bit lower quality | 37.2 GB | 40.2 GB | 64 GB |
On every Apple chip
The smallest memory size of each chip that runs Mistral Small 4 (119B), the best version that fits it, and the writing speed you can expect.
| Chip | Needs | Version | Tokens/s |
|---|---|---|---|
| M6 | no size fits | Too big | – |
| M5 Ultra | 256 GB | 8-bit Runs well | 78–139 |
| M5 Max | 128 GB | 4-bit Runs well | 75–133 |
| M5 Pro | 64 GB | 2-bit Runs well | 69–123 |
| M5 | no size fits | Too big | – |
| M4 Max | 128 GB | 4-bit Runs well | 66–118 |
| M4 Pro | 64 GB | 2-bit Runs well | 61–109 |
| M4 | no size fits | Too big | – |
| M3 Ultra | 256 GB | 8-bit Runs well | 53–95 |
| M3 Max | 128 GB | 4-bit Runs well | 49–86 |
| M3 Pro | no size fits | Too big | – |
| M3 | no size fits | Too big | – |
| M2 Ultra | 128 GB | 4-bit Runs well | 97–173 |
| M2 Max | 64 GB | 2-bit Runs well | 90–160 |
| M2 Pro | no size fits | Too big | – |
| M2 | no size fits | Too big | – |
| M1 Ultra | 128 GB | 4-bit Runs well | 97–173 |
| M1 Max | 64 GB | 2-bit Runs well | 90–160 |
| M1 Pro | no size fits | Too big | – |
| M1 | no size fits | Too big | – |
How to run it
- The easy way: download MLX Harness, free and open source. Search for Mistral Small 4 (119B), download the version that fits your Mac, and chat. It also serves the model to your other apps over an OpenAI-compatible API.Download MLX HarnessFree · Apple Silicon · macOS 14.5 or later
- In Terminal with Apple’s MLX:
pip install mlx-lmmlx_lm.chat --model mlx-community/Mistral-Small-4-119B-2603-4bitBuild: mlx-community/Mistral-Small-4-119B-2603-4bit on Hugging Face. Other precisions sit next to it.
Step by step: download and run LLMs on a Mac · how to use MLX Harness · MLX vs Ollama vs llama.cpp
About Mistral Small 4 (119B)
| Made by | Mistral AI |
|---|---|
| Parameters | 119B total · 6.5B active (mixture of experts) |
| Context | 256K tokens |
| Licence | Apache-2.0 |
| Released | 2026-01 |
| Input | Text and images |
| Intelligence | 11.3 on the Artificial Analysis Intelligence Index |
| Weights | mistralai/Mistral-Small-4-119B-2603 |
Questions
- How much memory does Mistral Small 4 (119B) need?
- About 66.9 GB of weights at 4-bit, plus room for the context: 72.3 GB in all. On a Mac that means 128 GB of unified memory or more, since macOS lets the GPU use about 75% of it by default.
- How fast is Mistral Small 4 (119B) on a Mac?
- Writing speed depends on memory bandwidth. The fastest Mac for it here is an M2 Ultra with 128 GB at about 97–173 tokens a second (4-bit). The smallest Mac that runs it, with 128 GB, manages about 97–173 on an M2 Ultra.
- Why does Mistral Small 4 (119B) run faster than its size suggests?
- It is a mixture-of-experts model: it holds 119B parameters in memory but reads only 6.5B for each token, so it writes at the speed of a much smaller model.
Similar models
For your model card or README
Shows the smallest Mac that runs Mistral Small 4 (119B) well and links here.
[](https://mlx.app/run/mistral-small-4)How it’s worked out
Memory: the model’s parameters at 8, 6, 4, 3 or 2 bits per weight plus MLX’s group scales (half a bit per weight), or its native format, with room for the context on top. By default macOS lets the GPU use about 75% of unified memory. Speed: memory bandwidth divided by the bytes read per token is the ceiling for one conversation; in llama.cpp’s public Apple Silicon benchmarks real runs reach about 45% to 80% of it, which is the range shown. These are estimates, not measurements: measure your own Mac on MLX Bench. Model figures from each official model card, chip figures from Apple’s Tech Specs. Checked 2026-10-19. Same data as mlx.app/compare.