Reflection AI boosts US ambition to compete with Chinese ‘open’ models
Reflection AI's Beam claims parity with DeepSeek on open weights—the US-China AI race just got a speed bump.
Why it matters
A US startup is releasing an open-weight model (Beam, 501B sparse MoE) that claims performance on par with leading Chinese competitors. This matters because open-weight releases shift the capability race away from proprietary walled gardens and raise the stakes for US AI sovereignty and cost-competitive inference.
The key facts
7 to knowReflection AI released Beam, a 501B parameter sparse mixture-of-experts model
Beam claims performance parity with leading Chinese 'open' models (likely DeepSeek reference)
Model released as open weights, not closed API
Framed as US response to Chinese AI competitiveness
Story filed under geopolitical competition lens, not pure capability benchmarking
No independent benchmark data provided in source; claim is startup assertion
Published October 5, 2026
Go to the source
Financial Times Technologyft.com
Publisher excerpt: Start-up claims new model Beam is on par with leading Chinese competitor as battle with Beijing intensifies