Memory at each precision
| Version | Weights | With context | Smallest Mac |
|---|---|---|---|
| 8-bit | 532 GB | 575 GB | more than any Mac |
| 6-bit | 407 GB | 440 GB | more than any Mac |
| 4-bit | 282 GB | 304 GB | 512 GB |
| 3-bit lower quality | 219 GB | 237 GB | 512 GB |
| 2-bit lower quality | 157 GB | 169 GB | 256 GB |
On every Apple chip
The smallest memory size of each chip that runs Beam 1.0, 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 | 512 GB | 4-bit Runs well | 42–74 |
| M5 Max | no size fits | Too big | – |
| M5 Pro | no size fits | Too big | – |
| M5 | no size fits | Too big | – |
| M4 Max | no size fits | Too big | – |
| M4 Pro | no size fits | Too big | – |
| M4 | no size fits | Too big | – |
| M3 Ultra | 512 GB | 4-bit Runs well | 29–51 |
| M3 Max | no size fits | Too big | – |
| M3 Pro | no size fits | Too big | – |
| M3 | no size fits | Too big | – |
| M2 Ultra | 192 GB | 2-bit Only just | 50–89 |
| M2 Max | no size fits | Too big | – |
| M2 Pro | no size fits | Too big | – |
| M2 | no size fits | Too big | – |
| M1 Ultra | no size fits | Too big | – |
| M1 Max | no size fits | Too big | – |
| 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 Beam 1.0, 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: There is no community MLX build yet, so make one from the official weights:
pip install mlx-lmmlx_lm.convert --hf-path reflection/Beam-1.0-501B-A23B -q
Step by step: download and run LLMs on a Mac · how to use MLX Harness · MLX vs Ollama vs llama.cpp
About Beam 1.0
| Made by | Reflection AI |
|---|---|
| Parameters | 501B total · 23B active (mixture of experts) |
| Context | 1M tokens |
| Licence | Apache-2.0 |
| Released | 2026-10 |
| Input | Text |
| Intelligence | 30.3 on the Artificial Analysis Intelligence Index |
| Weights | reflection/Beam-1.0-501B-A23B |
Questions
- How much memory does Beam 1.0 need?
- About 282 GB of weights at 4-bit, plus room for the context: 304 GB in all. On a Mac that means 512 GB of unified memory or more, since macOS lets the GPU use about 75% of it by default.
- How fast is Beam 1.0 on a Mac?
- Writing speed depends on memory bandwidth. The fastest Mac for it here is an M5 Ultra with 512 GB at about 42–74 tokens a second (4-bit). The smallest Mac that runs it, with 512 GB, manages about 42–74 on an M5 Ultra.
- Why does Beam 1.0 run faster than its size suggests?
- It is a mixture-of-experts model: it holds 501B parameters in memory but reads only 23B 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 Beam 1.0 well and links here.
[](https://mlx.app/run/beam-1.0)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.