• GPT-5 may launch in August, with “mini” and “nano” versions also planned.
• Model could merge AI tools, making ChatGPT simpler.

OpenAI may be just weeks away from releasing its next big AI model, GPT-5. According to a report by The Verge, the model could arrive as early as August. Sources familiar with the company’s plans say internal testing is nearly complete, and all signs point to the model being ready to launch soon.

The news follows earlier reports that Microsoft engineers were preparing server space for a possible GPT-5 release as of May. While that timeline shifted, the growing signs of internal use suggest that OpenAI is now in the final stages of rollout.

CEO Sam Altman has begun to publicly acknowledge the model. He posted on social media, “we are releasing GPT-5 soon,” and discussed the issue further in a podcast appearance with comedian Theo Von. During the show, Altman described asking GPT-5 a question he didn’t fully understand. “It answered it perfectly,” he said. “I felt useless relative to the AI.” He called the moment “weird,” but admitted it was a clear example of how far the technology has come.

What makes GPT-5 different?

Sources say GPT-5 will include more than an upgrade in power. It’s expected to merge multiple tools and abilities into one system. ChatGPT users often switch between different models, depending on the task – whether it’s writing, coding, browsing the web, or generating video.

GPT-5 could combine these demands into a single model. That means users might be able to use the same interface for all tasks.

Altman previously described GPT-5 as “a system that integrates a lot of our technology,” which supports the concept. Unlike past updates like GPT-4o or the o3 model, GPT-5 might be the first to amalgamate models.

The new model is also expected to launch with “mini” and “nano” versions. The smaller models may be available through OpenAI’s API and are designed to give developers and businesses flexible options – trading power for faster performance or lower cost.

Improvements developers are watching for

There are also hints that GPT-5 will be better at writing code. In a post earlier this month, Altman told a story many interpreted as a nod to improved coding skills. Major model updates usually come with better performance in programming and creative tasks, and GPT-5 is likely to follow that trend.

Another area to watch is the model’s memory, or “context window.” The term refers to how much information a model can process at once. The larger the window, the better it can stay on topic or remember earlier parts of a conversation. Developers have been asking for this feature to grow, and it’s likely that GPT-5 will offer expanded capacity.

GPT-5 has also been linked to OpenAI’s internal BioSec Benchmark repository, according to researchers cited by TechRadar. This suggests the model is already being tested for use in sensitive areas like biosecurity – an early sign of trust in its capabilities.

OpenAI research scientist Alexander Wei also mentioned the model recently, saying: “We are releasing GPT-5 soon.” While vague, it adds to the list of comments suggesting that a release is just around the corner.

A quiet lead-up compared to past launches

OpenAI has mostly kept details about GPT-5 under wraps. Unlike some earlier launches, where public demos and blog posts offered clues, this one has been relatively quiet. Aside from a few social media mentions and indirect references, the company has avoided giving too much away.

Still, anticipation is building. With competitors like Google’s Gemini and Anthropic’s Claude pushing hard to stay ahead in generative AI, GPT-5 is OpenAI’s answer to keep its place as one of the most widely used AI platforms.

A possible Sora update could be part of the launch

In addition to GPT-5, OpenAI may also be preparing an update to Sora, its video generation tool. Tech analyst Tibor Blaho recently shared a screenshot on social media that appeared to show internal filenames labeled “Sora 2,” leading many to believe a new version is in the works.

Sora was originally released in December 2024 and was seen as a big step forward for AI-generated video. But since then, other models like Google’s Veo 3, Kling 2, and Runway have started to gain more attention. A Sora update alongside GPT-5 would help OpenAI stay in the conversation in the fast-moving AI video space.

Though OpenAI hasn’t confirmed Sora 2, adding new features or performance upgrades would make sense – especially if the company is trying to show how its tools can work better together under one model family.

With just weeks to go, developers, researchers, and investors will be watching closely. If GPT-5 delivers on its early promise, it could mark a major step toward more unified, flexible AI systems built for a wider range of real-world use.