By Key Points
- OpenAI launches o3-pro for Pro and Team users
- o3-pro outperforms o3 and o1-pro in key domains
- Model priced 87% lower than o1-pro
- Open-weights model release delayed to late summer
OpenAI is rolling out its latest language model, OpenAI o3-pro, to all ChatGPT Pro and Team users starting this week. The model replaces the previous o1-pro and is available via the model picker in ChatGPT. Enterprise and Edu users will get access next week.
The company has also revealed major pricing updates and a delay in the much-anticipated open-weights model.
We’re cutting the price of o3 by 80% and introducing o3-pro in the API, which uses even more compute.
o3:
Input: $2 / 1M tokens
Output: $8 / 1M tokensNow in effect.
We optimized our inference stack that serves o3. Same exact model—just cheaper.https://t.co/LkObjZtg9s pic.twitter.com/CrzhChUYNx
— OpenAI Developers (@OpenAIDevs) June 10, 2025
OpenAI o3-pro brings top-tier performance and major price cuts
OpenAI describes o3-pro as a major step forward in both capabilities and cost efficiency. It builds on the underlying architecture of o3 but surpasses it in key areas. According to OpenAI:
OpenAI o3-pro has access to tools that make ChatGPT useful—it can search the web, analyze files, reason about visual inputs, use Python, personalize responses using memory, and more.
— OpenAI (@OpenAI) June 10, 2025
One of the standout achievements is the model’s passing of the rigorous 4/4 reliability benchmark. This means the model must answer a question correctly in four out of four attempts, ensuring high consistency—an essential factor for professional applications.
OpenAI o3-pro also continues the academic strengths of its predecessor. Like OpenAI o1-pro, the new model excels in math, science, and coding, as shown in various academic evaluations.
Another significant update is pricing. OpenAI has slashed the cost of the base o3 model by 80%:
-
o3 inputs now cost $2 per million tokens, and outputs $8 per million tokens.
-
o3-pro inputs are priced at $20 per million tokens, and outputs at $80 per million tokens—which is 87% cheaper than o1-pro.
The lower pricing opens new possibilities for users, particularly startups and small businesses, to experiment with advanced AI at a fraction of the previous cost.
In addition, o3-pro offers powerful tool integrations, including:
-
Web search
-
File analysis
-
Visual reasoning
-
Python execution
-
Personalised responses using memory
These tools significantly expand what users can achieve with ChatGPT.
Meanwhile, other tech giants are also pushing AI boundaries. NVIDIA recently warned the UK about the urgent need to upgrade its AI infrastructure, highlighting how competitive the global AI race has become.
OpenAI delays open-weights model for “amazing” new development
While OpenAI users welcomed the o3-pro rollout, CEO Sam Altman shared that the open-weights model release, initially expected this June, will be delayed until later this summer.
In a post on X, Altman explained that the delay is due to a new breakthrough from the research team. He described the development as:
o3-pro is rolling out now for all chatgpt pro users and in the api.
it is really smart! i didnt believe the win rates relative to o3 the first time i saw them.
— Sam Altman (@sama) June 10, 2025
The open-weights model is highly anticipated by the AI community, as it would provide transparency and greater flexibility for developers and researchers.
As AI development heats up, major players are advancing fast. Elon Musk’s xAI recently raised fresh capital through debt to scale its efforts. Meanwhile, Apple is integrating more AI features into its ecosystem, and even launching innovations like an AI-powered workout buddy in watchOS 26.
With OpenAI now focusing on refining its model further, users can expect a more advanced release when it arrives. The company also seems to be two stages away from achieving Artificial General Intelligence (AGI), with o3-pro marking another important milestone in this journey.
What's your reaction?
