OpenAI's Next-Gen AI: o3, o4-mini, and GPT-5 Launching Soon
After initially canceling the consumer launch of its o3 reasoning model in February, OpenAI has announced a change of plans. The company now intends to release both o3 and its next-generation successor, o4-mini, within the next few weeks.
According to a recent post on X by OpenAI CEO Sam Altman, this shift in strategy is closely tied to the development of GPT-5, which the company envisions as a unified model boasting enhanced reasoning capabilities.
"[W]e are going to be able to make GPT-5 much better than we originally thought," Altman stated. He also acknowledged the challenges in seamlessly integrating all the components, emphasizing the need to ensure sufficient capacity to handle the anticipated high demand. The rollout of GPT-5 is now expected in the coming months, slightly later than initially projected.
OpenAI has previously indicated that GPT-5 will offer unlimited chat access at the "standard intelligence setting," subject to certain abuse thresholds. Subscribers to ChatGPT Plus and ChatGPT Pro plans will have the option to run GPT-5 at progressively higher levels of intelligence.
Altman has also hinted at the integration of various features into GPT-5, including voice, Canvas, search, and deep research capabilities. The overarching goal is to create a unified model capable of utilizing all of OpenAI's tools, discerning when to engage in prolonged thinking, and providing utility across a wide spectrum of tasks.
OpenAI faces increasing competition from rivals like DeepSeek, a Chinese AI lab adopting an "open" approach to model releases. These competitors make their models available for experimentation and commercialization, contrasting with OpenAI's more controlled strategy. In addition to o3, o3 pro, o4-mini, and GPT-5, OpenAI plans to launch its first open language model since GPT-2 in the coming months. This model will feature reasoning capabilities and undergo rigorous safety evaluations.
Source: TechCrunch