A Brooklyn Startup Wants to Be the West's Open-Weights Answer to DeepSeek
Reflection's Beam claims to match Chinese open models on reasoning at a third of the inference cost — with $4.7 billion raised and an 'AI factory' pitch aimed at enterprises and governments.
The open-weights AI race has a new Western contender. Reflection AI, a two-year-old Brooklyn startup, has unveiled Beam, its first frontier model — and the pitch is direct: reasoning on par with China's leading open models, at a fraction of the cost.
The numbers
Beam is a text-only mixture-of-experts model: 501 billion total parameters with 23 billion active, pretrained on 23.8 trillion tokens, with a one-million-token context window. For comparison, Z.ai's GLM-5.2 runs roughly 744 billion total parameters with 40 billion active.
According to TechCrunch, Reflection says Beam scores on par with GLM-5.2 on advanced reasoning benchmarks and outperforms today's leading Western open models while using three to four times less inference compute. The company trained it with high-compute reinforcement learning tuned for reasoning, coding, and agentic work, describing Beam as a "workhorse model" for enterprises, the public sector, and developers. Those performance claims have not been independently verified, the report notes.
Reflection's own benchmarks also show Beam outscoring Inkling — the open model released in July by Mira Murati's Thinking Machines Lab — on four coding tests where both publish results. The caveat: Inkling is multimodal, while Beam is text-only.
The business behind the model
Reflection was founded in 2024 by two former Google DeepMind researchers and has raised roughly $4.7 billion from backers including Nvidia, Sequoia Capital, and Lightspeed Venture Partners, per PitchBook data cited by TechCrunch. Its last round carried a $25 billion pre-money valuation. The startup has also been locking up compute aggressively: this summer it signed deals collectively worth more than $7 billion with SpaceX and Nebius to secure Nvidia GB300 chips through 2029.
The real product, though, may be the pitch. Reflection is selling "AI factories" — letting enterprises and sovereign nations train its models on their own proprietary data to build customized, local AI systems. Nvidia CEO Jensen Huang, whose company backs Reflection, has championed the same concept. Axios previously reported that hedge funds and trading firms are among the eager customers, and Reflection says it has begun testing a sovereign AI factory partnership with Shinsegae Group in South Korea.
Open, with an asterisk
Reflection says it will release Beam's weights and full technical details this month, distributed through hyperscalers and neoclouds with open-source library integrations at launch. That openness is the strategic point: the startup is positioning against closed labs like Anthropic and OpenAI, against Chinese open models, and against Western players from Mistral to Meta.
Whether "Western answer to DeepSeek" is a marketing line or a technical achievement depends on what independent benchmarks show once the weights are actually out. Until then, Beam is a well-funded promise — one of the best-capitalized in the open-model field.
Written by the BidAsk Newsroom. Based on reporting from the sources below.
Which AI tool is actually worth your money?
ToolROI is a free AI-tool ROI calculator for freelancers — compare what tools like Jasper, Copy.ai, Kit, and Semrush really cost you versus what they earn back.
Calculate my ROI →Free tool run by the same publisher as BidAsk. No affiliate links.
