OpenAI Releases Its Most Powerful Model o1: 92.8% in PhD Physics, IOI Gold-Level Performance
OpenAI's new o1 model excels in complex reasoning, scoring 92.8% in PhD-level physics and surpassing 93% of programmers in IOI.
Today, OpenAI made waves by releasing the long-anticipated o1 model, codenamed "Strawberry."
The o1 series excels at complex reasoning but takes more time for each response.
While GPT-4o scored a mere 59.5% on PhD-level physics problems, o1 rocketed to an impressive 92.8%!
Yes, the legendary "Strawberry" is finally here.
This new model focuses on complex tasks, solving tougher problems in fields like science, programming, and math.
It works slower than GPT-4o and comes at a higher cost, but there’s also a smaller, more affordable version called o1-mini. No more waiting either:
Starting today, ChatGPT Plus and Team users can access both o1-preview and o1-mini. Business and education users will get access next week, and o1-mini will soon be available to all free users.
To use it, just select it manually from the model dropdown menu.
For now, o1-preview has a 30-message weekly limit, while o1-mini has 50.
Developers at tier 5 can also call these models via API but at a higher cost: $15 per million input tokens and $60 per million output tokens for o1-preview.
In comparison, GPT-4o costs just $5 and $15 for the same input and output tokens.
Keep reading with a 7-day free trial
Subscribe to AI Disruption to keep reading this post and get 7 days of free access to the full post archives.