Reflection AI Unveils 501 Billion-Parameter Beam, Claiming 3-4x Lower Compute Than Chinese Rivals
Updated
Updated · TechCrunch · Oct 5
Reflection AI Unveils 501 Billion-Parameter Beam, Claiming 3-4x Lower Compute Than Chinese Rivals
3 articles · Updated · TechCrunch · Oct 5
Summary
Beam is Reflection AI’s first frontier open-weight model, a text-only system with 501 billion parameters, 23 billion active parameters and a 1 million-token context window.
Reflection says the model matches Z.ai’s GLM-5.2 on advanced reasoning benchmarks while using 3-4x less inference compute, though those performance claims have not been independently verified.
The Brooklyn startup is pitching Beam as a lower-cost Western alternative to Chinese open models and to closed systems from OpenAI and Anthropic, targeting enterprises, public-sector users and developers.
Reflection has raised about $4.7 billion at a $25 billion pre-money valuation and this summer signed more than $7 billion in SpaceX and Nebius deals to secure Nvidia GB300 chip access through 2029.
Weights and full technical details are due later this month as Reflection pushes its broader "AI factory" strategy, including sovereign and enterprise deployments such as a pilot with South Korea’s Shinsegae Group.