OpenAI has announced that the release of its new open AI model will be delayed until later this summer, rather than coming out in June as originally planned.
CEO Sam Altman shared this news on , saying the team made an unexpected breakthrough that was exciting but needed more time to finish properly.
He assured that the wait will be worth it.
This new open model is designed to match the reasoning power of OpenAI’s current “” models. OpenAI is aiming to make it the best-performing open model available, outshining competitors like .
“Open models” are AI systems that anyone can view, use, and build upon, a contrast to the more closed and private models that companies usually keep for themselves.
However, the AI field has become much more competitive lately. For instance, a company called Mistral just released its own new set of open AI models named “Magistral.”
Meanwhile, in April, a Chinese company called Qwen launched hybrid AI models that can either take time to think carefully or respond quickly, depending on the situation.
To stay ahead in this fast-changing space, OpenAI may include advanced features in its upcoming model.
One idea being considered is letting the open model connect to OpenAI’s more powerful cloud-based systems to answer difficult questions.
However, it’s still uncertain if that feature will make it into the final product.
This release is also important for OpenAI’s public image. In the past, the company has been criticized for not sharing its models openly with the research community.
Altman himself admitted that OpenAI might have made the wrong call by keeping things too closed.
Now, by releasing a high-quality open model, OpenAI hopes to rebuild trust and prove that it can lead in both innovation and openness.
Although the model is delayed, OpenAI is promising something special, and it wants to make sure it gets it right.
Thoughts on this delay? Do you think it’s okay to delay important releases if it means the product arrives in a better shape? Tell us below in the comments, or reach us via our or .
