Reflection unveils open-weight model Beam, saying it excels at coding and agentic tasks, uses 3x-4x less compute than comparable models, and nears Qwen 3.8-Max (Semafor)

Reflection AI has unveiled its first open-weight model, Beam, which it claims excels at coding and agentic tasks. According to Reflection AI, Beam uses 3x-4x less compute than comparable models.

Beam is a 501-billion-parameter model with 23 billion active parameters, pre-trained on 23.8 trillion tokens and has a 1 million token context window. TechCrunch reports that Reflection AI claims Beam matches the performance of leading Chinese open models on advanced reasoning benchmarks at dramatically lower costs.

Reflection AI positions Beam as a "workhorse model" for enterprises, the public sector, and developers. The company's performance claims have not been independently verified. According to Semafor, Beam nears the performance of Qwen 3.8-Max.

Brief written by urgent.news from Techmeme, TechCrunch — 2 reports on this story. Machine-written — may contain errors; check the original before relying on it.