Reflection AI releases first open model to rival China

The open-source startup Reflection AI released its first open-weight model Monday, stating it rivals leading open technologies made in the U.S. and China.

Reflection AI, which has the backing of tech giant Nvidia, said in a blog post Monday its Beam model is on par with some leading open-source models made in China, including Chinese AI startup Z.ai’s GLM-5.2 model, while it’s close to Alibaba’s Qwen3.8-Max.

Reflection, founded by two former Google DeepMind researchers, described Beam as a “powerful workhorse model,” adding it was designed for coding and agentic tasks and reasoning to break down complex problems.

BestReviews is reader-supported and may earn an affiliate commission.

The release comes amid concerns over China’s massive build-out of open-source technologies, which are then sold to people and enterprises around the world for cheaper rates. But with cheap costs also come potentially less secure models, prompting scrutiny around national security implications.

Unlike private models hosted by firms like OpenAI or Anthropic, open-source systems are often smaller or more fine-tuned and live in the public domain. This gives virtually anyone or any company access to download and customize the model for a wide range of uses.

In some cases, a model may not be open source but can have open weights, meaning the ways a model is trained to sift through information and formulate answers are made public.

According to research from venture capital firm Andreessen Horowitz, 80 percent of developers around the world that use open-source tools are building with Chinese models.

These companies include U.S.-based firms. Last year, Airbnb CEO Brian Chesky said the company was “relying a lot” on Alibaba’s Qwen model and described it as “very good” and “fast and cheap.”

Beam is getting its final evaluations, and the weights and technical details will be released later this month, Reflection AI said.