Performance Improvement - Performance improved from the 49th to the 98th percentile in IOI in one year [1] - The improvement was achieved without training any specialized models [1] - The same Reinforcement Learning (RL) was used as for everything else [1]

Model Performance Improvement - OpenAI's models have significantly improved in competitive programming skills, achieving IOI gold-level performance in just 2.5 years [1] - Initially, GPT-4 struggled with simple Codeforces problems and faced "oom-ing" issues [1] Project Context - The project involved teaching models to reason and use tools by enhancing their competitive programming abilities [1] Team Contribution - The progress reflects the hard work of many individuals within OpenAI [1]

RT Sheryl Hsu (@SherylHsu02)1/n I’m thrilled to share that our @OpenAI reasoning system scored high enough to achieve gold 🥇🥇 in one of the world’s top programming competitions - the 2025 International Olympiad in Informatics (IOI) - placing first among AI participants! 👨‍💻👨‍💻 https://t.co/k3RQxIzXPH ...

Product & Pricing Strategy - Company is considering offering a small number of GPT-5 Pro queries monthly to Plus subscribers [1] - Individual user willing to pay $1k/month for GPT-5 Pro with 2x input tokens [1] - Company might consider offering 2x input tokens for $1k/month [1] Product Performance - GPT-5 Pro is considered "ridiculously good" by at least one user [1]

If you have been following the GPT-5 rollout, one thing you might be noticing is how much of an attachment some people have to specific AI models. It feels different and stronger than the kinds of attachment people have had to previous kinds of technology (and so suddenly deprecating old models that users depended on in their workflows was a mistake).This is something we’ve been closely tracking for the past year or so but still hasn’t gotten much mainstream attention (other than when we released an update ...

User Engagement - The percentage of free users utilizing reasoning models daily has increased significantly from less than 1% to 7% [1] - The percentage of plus users utilizing reasoning models daily has increased significantly from 7% to 24% [1] Future Expectations - The company anticipates a substantial increase in reasoning model usage over time [1] Recommendation - Rate limit increases are deemed important to accommodate the expected growth in reasoning model usage [1]

Model Updates - Significantly increasing rate limits for reasoning for ChatGPT Plus users [1] - All model-class limits will shortly be higher than they were before GPT-5 [1] - A UI change will shortly indicate which model is working [1]

Capacity Allocation Strategy - The company plans to share its capacity tradeoff strategy in the coming months [1] - The strategy will address capacity allocation between different products and user groups [1] Capacity Tradeoffs Considerations - Capacity tradeoffs will consider ChatGPT vs the API [1] - Capacity tradeoffs will consider existing users vs new users [1] - Capacity tradeoffs will consider research vs product [1]

Wanted to provide more updates on the GPT-5 rollout and changes we are making heading into the weekend.1. We for sure underestimated how much some of the things that people like in GPT-4o matter to them, even if GPT-5 performs better in most ways.2. Users have very different opinions on the relative strength of GPT-4o vs GPT-5 (just the chat model, not the advanced reasoning one). This is a cool thing you can try: https://t.co/O9Nu6LPC7f3. Long-term, this has reinforced that we really need good ways for dif ...

GPT-5 rollout updates:*We are going to double GPT-5 rate limits for ChatGPT Plus users as we finish rollout.*We will let Plus users choose to continue to use 4o. We will watch usage as we think about how long to offer legacy models for.*GPT-5 will seem smarter starting today. Yesterday, the autoswitcher broke and was out of commission for a chunk of the day, and the result was GPT-5 seemed way dumber. Also, we are making some interventions to how the decision boundary works that should help you get the righ ...