
FaceTime Like a Pro
Get our exclusive Ultimate FaceTime Guide 📚 — absolutely FREE when you sign up for our newsletter below.
FaceTime Like a Pro
Get our exclusive Ultimate FaceTime Guide 📚 — absolutely FREE when you sign up for our newsletter below.
OpenAI teases GPT-5 with a ‘LIVE5TREAM’ event on August 7, promising faster responses, smarter model switching, and improved reasoning.
OpenAI is all but confirming the arrival of GPT-5, teasing a major announcement this Thursday with a not-so-subtle nod to its next flagship model.
The company posted on X (formerly Twitter) that it will host a “LIVE5TREAM” event on August 7 at 10AM PT. The stylized title with a “5” in place of the “S” isn’t exactly cryptic, and all indications point to GPT-5 finally stepping into view.
OpenAI hasn’t kept things low-key in the run-up. On Sunday, CEO Sam Altman posted a screenshot that showed “ChatGPT 5” selected within the ChatGPT interface, a quiet but direct signal that the model is already in testing.
The next day, Boris Power, OpenAI’s head of applied research, chimed in with: “excited to see how the public receives GPT-5!” The company hasn’t issued a formal statement, but the message is clear enough.
The Verge’s Tom Warren previously reported that OpenAI was aiming for an early August rollout, with Microsoft provisioning server capacity ahead of time. With this week’s teaser, that timeline appears to be holding.
Altman had also prepped the announcement with a cryptic post last week, writing: “something big-but-small today, and then a big upgrade later this week.”
That smaller launch was gpt-oss: two open-weight language models designed for local use, including one gpt-oss-20b optimized specifically for Apple Silicon Macs. Now the focus shifts to GPT-5, the company’s first major model refresh since the iterative rollout of GPT-4 variants like 4.1, 4.5, and GPT-4o.
Sam Altman has already hinted at the model’s capabilities, noting in a July podcast that GPT-5 rewrote a confusing email so well that he felt “useless” by comparison. While OpenAI hasn’t detailed what’s new, some expected features are beginning to surface.
Early signs suggest faster performance, smarter model routing, and more efficient processing. One possible change: ChatGPT may soon auto-select the best model version for a given task, rather than requiring users to choose manually.
Another rumored feature is step-by-step “chain-of-thought” visibility, giving users a clearer window into how the model reasons through prompts. Altman has also mentioned a long-term goal of unifying OpenAI’s different model tracks, such as the GPT-4o line, into a more cohesive user experience.
Still, rollout concerns remain. Altman has previously said GPT-5 was delayed due to technical challenges and server strain. And even this week, he warned the launch could be “slightly choppy” as demand ramps up.
Despite advances from rivals Google with Gemini 2.5 and Anthropic’s Claude 4 ChatGPT remains the most widely used consumer AI product. If GPT-5 delivers the kinds of improvements Altman has hinted at, it could reinforce OpenAI’s lead in a crowded, fast-moving field.
The livestream kicks off Thursday at 10AM PT / 1PM ET, presumably on OpenAI’s YouTube channel. For users, developers, and competitors alike, this may be the most consequential update ChatGPT has seen in over a year.
Are you watching the launch? Curious what GPT-5 will change for you? Drop your thoughts in the comments.
Also Read: