#MarketPullback 🤖 OpenAI announced new language models o3 and o3-mini that promise to revolutionize the industry. Here’s why this event is so important:
Why is o3 better?
🔴 30% smarter in coding compared to o1.
🔴 Solves mathematical problems 15% more accurately.
🔴 Capable of explaining its reasoning and providing highly accurate answers.
What’s special?
🔴 The model was trained to 'think' before responding. It first recounts its reasoning chain and then gives the conclusion.
🔴 Users will be able to choose how much time to give the model for analysis: more time means more accurate results.
Benchmarks are impressive:
🔴 In the ARC-AGI test, the model achieved 87.5%. This is nearly AGI (Artificial General Intelligence) level.
🔴 In comparison, a result of 85% is considered human level.
When to expect?
🔴 o3-mini will be released in January 2025.
🔴 After that, the company plans to release the full version of o3.
The model is so good that coders may seriously rethink their role in the future. OpenAI is making a big step towards creating AI that can replace many professions. OpenAI CEO Sam Altman joked that the number o2 was skipped 'out of respect for Telefonica' and the tradition of coming up with strange names