Reflection has released Beam, its first open-weight model. The mixture-of-experts system activates just 23 billion of its 501 billion parameters per token and aims to match GLM 5.2 on coding and reasoning while using three to four times less compute. The strategy against Chinese open-weight rivals like Deepseek and Qwen is efficiency over raw performance.
The article Reflection's Beam becomes the most capable open-weight model built outside China appeared first on The Decoder.
source & further reading
the-decoder.com — original article
Cohere pitches North 2 as the enterprise AI control room that works with any model
Reka AI's omni-model Rho-1 handles text, images, video, and robot control in a single model
Google researchers find a way to keep self-improving AI agents from memorizing their tests