After a brief pause in February, OpenAI has officially announced the upcoming release of its o3 reasoning model, along with a next-gen variant named o4-mini. According to a recent update by Sam Altman, CEO of OpenAI, both models are expected to launch “in a couple of weeks.”
This shift comes as the company works to perfect GPT-5 — the anticipated unified model that integrates reasoning, memory, and advanced tool use. Altman stated that while the new model’s performance is exceeding expectations, the integration process has proven more difficult than anticipated.
“We are going to be able to make GPT-5 much better than we originally thought… but we also want to ensure we have enough infrastructure to handle the demand,” Altman posted on X.
What’s Coming with GPT-5?
Once released in the coming months, GPT-5 will bring:
- Unified multimodal capabilities (text, voice, canvas, search, and research).
- Tiered access: from standard users to ChatGPT Plus and Pro subscribers, each getting access to increasingly powerful versions of the model.
- A new open-source model — the company’s first since GPT-2 — offering reasoning capabilities with rigorous safety controls.
Rising Pressure from Global Competitors
OpenAI’s strategic pause and focus on safety contrasts with newer players like DeepSeek, a Chinese AI lab that’s gaining attention for its open model releases. These models are more readily available for experimentation and even commercialization pressuring OpenAI to match speed with precision.
While OpenAI plans its releases cautiously, its aim remains the same: to build general-purpose AI systems that can assist in a broad spectrum of tasks from quick chats to long-form research.
Also Read : Devin AI Launches Affordable Pay-As-You-Go Plan Amid Rising Popularity
