OpenAI Teases The Next Big Thing GPT 5
OpenAI has had a packed schedule this year, rolling out a series of impressive AI tools. We've seen everything from Operator, its first AI agent, to the new 4o image generator and the o3 reasoning chatbot integrated into ChatGPT. The most recent launch, the upgraded ChatGPT Agent, was met with such overwhelming demand from Pro subscribers that OpenAI had to delay its availability for Plus and Team users.
Even with this flurry of activity, the AI community is holding its breath for what are considered the year's most significant updates. The spotlight is on the GPT-5 model, anticipated for a summer release, and a long-awaited open-source version of ChatGPT that has faced multiple postponements.
GPT-5 is generating the most excitement as it promises to revolutionize the entire chatbot experience, likely succeeding the current GPT-4o model that powers most ChatGPT interactions. A recent increase in teasers and leaks suggests that the next generation is just around the corner.
Sam Altman's Mysterious Coding Tool
Last Saturday, shortly after the ChatGPT Agent launch, OpenAI CEO Sam Altman shared an intriguing post on X. "Woke up early on a Saturday to have a couple of hours to try using our new model for a little coding project," he wrote. "Done in 5 minutes. It is very, very good. Not sure how I feel about it..."
While not an explicit confirmation, the post strongly hints at GPT-5. Advanced coding is an expected hallmark of the new model, and it's logical for the CEO to be testing it ahead of a public launch, especially since he previously mentioned a summer release window.
Adding to the buzz, prominent AI watcher Tibor Blaho highlighted a post from OpenAI engineer Alexander Wei. In an X thread about a research model that achieved a gold medal at the International Math Olympiad (IMO), Wei dropped a teaser: "btw, we are releasing GPT-5 soon, and we're excited for you to try it." He clarified, however, that the powerful math model was an experimental project and not slated for release for several months.
GPT-5 Spotted in Testing
Known for uncovering details on unreleased AI, Blaho followed up with more concrete evidence. Before Altman's tweet, Blaho discovered signs of GPT-5 being actively tested. A model named "openai/gpt-5-reasoning-alpha-2025-07-13" has been identified in testing logs. The name suggests it's an alpha version focused on reasoning, much like the existing ChatGPT o3 and o4-mini models.
This aligns with speculation that GPT-5 could be the first model that doesn't require users to select a specific mode. The AI may be intelligent enough to automatically engage its reasoning capabilities when a task demands it. To achieve this, OpenAI might be developing a suite of specialized models that operate under the unified GPT-5 banner.
Hours after Altman's post, OpenAI's official X account provided further clarification. "To be clear: We're releasing GPT-5 soon, but the model we used at IMO is a separate experimental model," the company stated. "It uses new research techniques that will show up in future models—but we don't plan to release a model with this level of capability for many months." While this manages expectations, the consistent message is clear: GPT-5 is on its way, though an exact timeline remains unconfirmed.