OpenAI reportedly plans one or two new models before year end, but its next flagship still needs alignment work
According to The Information, the flagship uses the same looped transformer technique as Astra, which makes safety monitoring harder.
OpenAI employees expect the company to release "one or two new models" before the end of the year, The Decoder reports, citing The Information. It is unclear whether the major upgraded model OpenAI is working on will be one of them. According to the report, that flagship uses the same looped transformer technique as Astra. The approach improves performance in math and coding but makes safety monitoring harder. The model still needs alignment work so that its goals match what OpenAI's researchers and customers expect. The company also has to improve the testing infrastructure around it before it can move forward. Once leadership is satisfied with the safety evaluations, there is another step. Under an executive order from President Trump on evaluating advanced AI systems, OpenAI has agreed to give the US government access to new models up to 30 days before outside partners get them, and longer before a public launch. The report suggests that safety readiness and government review now shape release timing at frontier labs. It comes after recent incidents involving model behavior at several developers.