Reflection AI released Beam on October 5, saying it matches leading Chinese open models on advanced reasoning benchmarks at dramatically lower costs. It is the company's first frontier model since its 2024 founding.
Reflection reported Beam scores on par with Z.ai’s GLM-5.2 on advanced reasoning benchmarks, measured on tasks like coding and reasoning. That compares with today’s leading Western open models, which it claims Beam outperforms.
Beam is built on high-compute reinforcement learning and targets enterprise, public sector, and developer use cases. Availability begins this month, initially for hyperscalers and neoclouds.
"Beam is a workhorse model for enterprises, the public sector, and developers," said Reflection AI. The company aims to provide AI factories that let institutions build customized, local AI systems by training its models on proprietary data.
The announcement follows reports that Reflection was close to a launch, confirmed by its Monday blog post. The company is positioning itself against closed labs like Anthropic and OpenAI, as well as Chinese open models.
Reflection did not say when Beam’s weights and full technical details will be released, and the company raised questions about independent verification of its performance claims. It plans to release Beam this month with integrations across open source libraries.
Source: techcrunch
