Beam 1.0
Beam 1.0 publishes its weights, but at 501B parameters (23.4B active per token) no consumer machine holds the checkpoint. This page covers the access paths that work and the open models that actually run locally.
- Parameters
- 501B(23.4B active)
- Format
- API
- Runs locally
- No
- Best for
- Agentic coding, Reasoning, Long context
Why you can't run Beam 1.0 locally
At 501B parameters (23.4B active per token), a Q4-class build of Beam 1.0 would need roughly 301 GB, only a maxed-out 512GB Mac Studio could even hold the weights, with no real headroom. That figure is arithmetic, not opinion: 0.6 GB per billion parameters is the standard Q4 rule we apply across the whole catalog. The weights themselves are public (official source linked below); capacity, not licensing, is the wall.
What works instead: the hosted API. For most workloads, the open models below deliver the same job on hardware that fits under a desk.
Reflection's open-weight model (Oct 2026, Apache 2.0): a text-only MoE with 501B total and 23.4B active parameters, 52 layers mixing gated GQA and sliding-window attention, and a 1M-token context window. Per Reflection's model card (BF16): SWE-bench Verified 81.6% and Terminal Bench v2.1 80.1%. BF16, FP8 and NVFP4 checkpoints are on Hugging Face and mainline llama.cpp supports it; most users will reach it through ollama run beam:cloud on Ollama Cloud.
What to run locally instead of Beam 1.0
| Model | Params | Runs on | Use |
|---|---|---|---|
| Qwen3 235B A22B | 235B | from 192 GB machines | quality |
| Qwen3.5 122B-A10B Instruct | 122B | from 96 GB machines | frontier-level reasoning |
| DeepSeek-R1 671B | 671B | from 512 GB machines | reasoning |
Get told when a better model fits your MacBook Air M5 16 GB
Get one email when a new open-weight model is a better fit for this machine.
For: MacBook Air M5 16 GB
You'll also get the weekly ModelFit email. By signing up you agree to our Privacy Policy. Unsubscribe from either email anytime, in one click.
Frequently asked questions
Can I run Beam 1.0 locally?
Not on hardware you can buy. The weights are public, but Beam 1.0 is a 501B-parameter model (23.4B active), and a Q4-class build would need roughly 301 GB, only a maxed-out 512GB Mac Studio could even hold the weights, with no real headroom. Until the ecosystem ships a smaller official build, the API or a smaller open model is the practical path.
How do I access Beam 1.0?
Through the vendor's hosted API. The official source is linked on this page.
What is the best local alternative to Beam 1.0?
Qwen3 235B A22B is the strongest local model we track (235B, from 192 GB machines), with Qwen3.5 122B-A10B Instruct close behind. Both run on a single high-memory Mac or GPU, see their pages for exact hardware.
Cite this page
ModelFit: Beam 1.0, specs, memory math and hardware verdicts. https://modelfit.io/models/beam-1.0/ (dataset updated 2026-10-24, CC BY 4.0).