Reflection debuts Beam, an open-weight AI model to rival Chinese models at lower compute cost

Reflection AI Inc. today introduced Beam, an open-source large language model with 501 billion parameters. The launch comes a few months after the startup raised funding at a $25 billion valuation. Around the same time, Reflection AI reportedly inked a $6.3 billion deal with SpaceX Corp. to rent Nvidia GB300 NVL72 appliances. It used those…

Reflection is aiming Beam and future models at enterprises and sovereign nations. The pitch is to build “AI factories,” a product that would let institutions build their own customized, local AI system by training Reflection’s AI models on their own proprietary data.

Western AI labs are finally trying to compete with China's DeepSeek and Zhipu (Z.ai) when it comes to efficient open-weight models. As a case in point, look no further than Reflection's Beam, a 501-billion-parameter, sparse Mixture of Experts (MoE) model that is insanely efficient.