# The Decoder — Reflection's Beam becomes the most capable open-weight model built outside China

- Company: The Decoder (the-decoder.com)
- Announced: 2026-10-06T12:11:22+00:00
- Category: open-weight-release
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://the-decoder.com/reflections-beam-becomes-the-most-capable-open-weight-model-built-outside-china/
- Record: https://forck.live/items/16192-reflection-s-beam-becomes-the-most-capable-open-weight-model-built-outside-china
- Subject: Reporting
- Open weights: yes
- Models affected: Beam
- License: Apache 2.0

Reflection released Beam, an open-weight mixture-of-experts model for coding and reasoning that activates 23 billion of its 501 billion parameters per token. The model matches GLM 5.2 on key benchmarks while using three to four times less compute, according to the company. Beam will ship later this month under the Apache 2.0 license.

## Evidence

Verbatim from https://the-decoder.com/reflections-beam-becomes-the-most-capable-open-weight-model-built-outside-china/:

> The mixture-of-experts model activates 23 billion of its 501 billion total parameters per token, keeping compute costs relatively low. On demanding reasoning tasks, Beam matches GLM 5.2 while using three to four times less compute, according to Reflection.

---

Record: https://forck.live/items/16192-reflection-s-beam-becomes-the-most-capable-open-weight-model-built-outside-china
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md