Reflection AI debuts Beam, an open-weight model aimed at rivalling China's best

Reflection AI has officially unveiled Beam, its first frontier open-weight AI model, positioning it as a Western answer to DeepSeek, Qwen and Z.ai, TechCrunch reported on Monday TechCrunch.

The two-year-old Brooklyn startup claims Beam matches the performance of leading Chinese open models on advanced reasoning benchmarks “at dramatically lower costs” — a claim that has not been independently verified. Reflection detailed the model in a lengthy blog post on Monday, describing Beam as a text-only mixture-of-experts model trained with high-compute reinforcement learning for reasoning, coding and agentic tasks TechCrunch.

Beam has 501 billion parameters, of which 23 billion are active, was pre-trained on 23.8 trillion tokens and offers a 1 million token context window, according to the company. For comparison, Z.ai’s GLM-5.2 carries roughly 744 billion total parameters with 40 billion active TechCrunch.

Reflection says Beam scores on par with GLM-5.2 on advanced reasoning benchmarks while using “3-4x less inference compute”, and that its own benchmarks show Beam outscoring Inkling — the open model from Mira Murati’s Thinking Machines Lab released in July — on four coding tests, though Inkling is multimodal and Beam is text-only TechCrunch.

The announcement confirms weekend reporting from Axios that the startup was close to launch Axios via TechCrunch. Reflection, founded in 2024 by two former Google DeepMind researchers, has raised roughly $4.7 billion from backers including Nvidia, Sequoia Capital and Lightspeed Venture Partners, per PitchBook, and was last valued at a $25 billion pre-money valuation TechCrunch.

This summer the startup signed deals collectively worth more than $7 billion with SpaceX and Nebius to secure access to Nvidia’s GB300 chips through 2029. Reflection says it will release Beam’s weights and full technical details this month, with distribution through hyperscalers and neoclouds TechCrunch.