OpenAI has delayed the release of its much-anticipated open model, now set for later this summer. CEO Sam Altman shared the update on X, explaining that unexpected yet impressive progress has meant the team needs extra time to polish the innovation. “We are going to take a little more time with our open-weights model, i.e. expect it later this summer but not June,” he noted, promising that the wait will be well worth it.
Originally scheduled for an early summer launch, the new model is designed to boost reasoning capabilities similar to OpenAI’s existing o-series. The company is aiming to outdo competitors like DeepSeek’s R1, even as the landscape shifts rapidly. Just this Tuesday, Mistral rolled out its Magistral series, and in April, Chinese firm Qwen introduced hybrid reasoning models that deftly balance deep problem-solving with quick, efficient responses.
OpenAI isn’t just focused on raw performance. The team is weighing potential integrations with its cloud-hosted AI models to handle complex queries—a move that could further enhance the model’s capabilities. While it’s not yet clear if these advanced features will make the final cut, this extra development time underscores the company’s commitment to quality.
This delay is particularly significant for researchers and developers who have been eagerly awaiting a more open approach. With previous criticisms around its open-sourcing policies, OpenAI appears determined to deliver a model that truly stands up to the best open solutions available.
