OpenAI to Launch o3, o4-mini and GPT-5 Amid Strategy Shift

615

After shelving the consumer launch of its o3 reasoning model earlier this year, OpenAI has announced that both o3 and its successor, o4-mini, are now scheduled for release in the coming weeks.

Reversal Tied to Upcoming GPT-5 Launch

In a post on X on Friday, OpenAI CEO Sam Altman attributed the change in strategy to the development of the company’s next flagship model, GPT-5.

According to Altman, GPT-5 is proving more promising than initially anticipated, though integrating its wide-ranging capabilities has proven challenging.

“We are going to be able to make GPT-5 much better than we originally thought,” Altman wrote.

“We also found it harder than we thought it was going to be to smoothly integrate everything.

“And we want to make sure we have enough capacity to support what we expect to be unprecedented demand.”

GPT-5 Rollout Expected Within Months

OpenAI now plans to launch GPT-5 “in a few months,” later than previously expected.

The new model will reportedly offer integrated access to several tools, including:

  • Voice
  • Canvas (visual interface)
  • Search
  • Deep research capabilities

The overarching goal, according to Altman, is to build a unified system that intelligently uses the full suite of OpenAI tools, adapts its reasoning time based on task complexity, and delivers general-purpose utility.

Tiered Access Levels for GPT-5

Once released, GPT-5 will be available at different tiers of access:

  • Standard intelligence level: Unlimited access, with usage monitored for abuse
  • Higher intelligence level: Available to ChatGPT Plus subscribers
  • Even higher intelligence level: Reserved for ChatGPT Pro users

New Reasoning Models on the Horizon

In addition to o3 and o4-mini, OpenAI also plans to launch:

o3 pro: An enhanced version of o3

A new open language model: The company’s first since GPT-2, featuring reasoning abilities and undergoing additional safety reviews

Pressure Mounts from Open-Source Competitors

OpenAI’s evolving strategy comes amid growing competition from labs such as China’s DeepSeek, which have embraced a more “open” model development approach.

These competitors are gaining traction by making their models freely available for research, testing, and in some cases, commercial use — in stark contrast to OpenAI’s more controlled rollout methods.

 

Comments are closed.