OpenAI has rolled out a significant wave of updates today, headlined by the launch of its most powerful AI model yet — o3 Pro — and a steep 80% price cut for the standard o3 model. These developments signal a major strategic push to dominate the market in both premium and accessible AI services.
What’s New with o3 Pro?
-
o3 Pro is now available to ChatGPT Pro and Team users, replacing the earlier o1 Pro model.
-
Though based on the same architecture as o3, it’s specially optimized for high-stakes tasks that demand deep reasoning, accuracy, and clarity — such as programming, science, education, and mathematics.
-
OpenAI recommends o3 Pro for users who value reliability over speed.
-
Equipped with advanced tools:
-
Web browsing
-
Python code execution
-
Image interpretation
-
Document analysis
-
Personalized memory
-
However, a few limitations currently exist:
-
Temporary chats are disabled due to technical issues.
-
Image generation and the Canvas interface are not yet available.
-
Users who need image output are advised to use GPT-4o, o3, or o4-mini.
OpenAI says Enterprise and Edu customers will gain access to o3 Pro next week.
80% Price Drop for o3 Model
OpenAI has also slashed o3 API pricing, making it far more accessible:
-
Input tokens: $2 per million (down from $10)
-
Output tokens: $8 per million (down from $40)
-
Cached tokens: additional discounts available
This drastic reduction positions OpenAI more competitively:
-
Google Gemini 2.5 Pro: $1.25–$2.50 input, up to $15 output
-
Anthropic Claude Opus 4: $15 input, $75 output
-
DeepSeek Chat/Reasoner: as low as $0.035 (cached input during off-peak)
Sam Altman posted on X:
“we dropped the price of o3 by 80%!! excited to see what people will do with it now. think you’ll also be happy with o3-pro pricing for the performance :)”
Open-Source Model Delayed
The anticipated open-source model from OpenAI, originally due in June 2025, has been delayed to later this summer.
-
According to Altman, the team made a "surprise discovery" that is “unexpected and quite amazing,” requiring more refinement before release.
-
The upcoming model is expected to challenge top open-access LLMs like DeepSeek R1, setting a new bar in the open-source landscape.
Environmental Impact Disclosures
Altman also addressed growing concerns around energy and resource use:
-
Energy per query: ~0.34 watt-hours(about what a light bulb uses in a few minutes)
-
Water per query: ~0.000085 gallons(about 1/15th of a teaspoon)
He reiterated OpenAI’s goal that "the cost of intelligence should eventually converge to the cost of electricity.”
Summary
OpenAI’s latest announcements reshape both premium AI usage with o3 Pro and developer access through massive cost reductions. The delay of the open-source model adds suspense, but also suggests something substantial is coming.
These moves intensify competition across the AI space — from Google’s Gemini to Anthropic’s Claude — while reinforcing OpenAI’s twin mission: advancing capability and democratizing access.
