Reflection AI, the Nvidia-backed startup founded by two former Google DeepMind researchers, is preparing to release its first open-weight model, Axios reported on October 4, 2026. The model is positioned as an American answer to the Chinese open-weight systems that currently dominate the space — DeepSeek and Alibaba's Qwen — which lead open-weight leaderboards on coding, math and cost efficiency. One important caveat: as of October 5, no model name, weights, license or benchmarks have been published — this is a reported release, not a launch, and Reflection declined to comment.

According to the report, the model is expected to trail the most advanced closed US models at first, while competing with the top open-weight models from China. Open-weight models let developers download and modify the model itself within its license terms — something that is not possible with the closed systems behind ChatGPT and Claude. Reflection has reportedly told interested parties in Washington that enterprises will be able to download and customize the model with their own private data, pitching a larger vision the startup calls the "AI factory": combining a company's own data, Reflection's models and its own compute to build cheaper, customized AI systems. Banks and the Pentagon are among the big AI users wary of Chinese models over security concerns, and Reflection is trying to fill that gap.

The context shows how much money is riding on this. Reflection was founded in March 2024 by Misha Laskin and Ioannis Antonoglou and, according to press reports, raised $2 billion in October 2025, with later reported valuations up to about $25 billion. Reports also cite more than $7 billion in compute commitments through 2029, including a Nebius partnership worth over $1 billion and monthly payments of about $150 million to SpaceX's Colossus cluster. Treat those figures as reported, not confirmed — none come from a filing. In March, the startup signed a memorandum of understanding with South Korea's Shinsegae Group to build a 250-megawatt AI factory in Korea running on Reflection's models and Nvidia chips.

Sources