# o3 and o4-mini

Source: https://openai.com/index/introducing-o3-and-o4-mini
Kiwi News: https://news.kiwistand.com/stories/o3-and-o4-mini?index=0x67ffe6c98ea257e2dceaef020cd2b8dc6336183a2e8c76af3c7d32db2fb384ebcfde8a66
Submitted by tim on 2025-04-16 UTC, 2 upvotes, 3 comments

## Summary

OpenAI released o3 and o4-mini, its smartest models to date, which can agentically use and combine every tool within ChatGPT, including web search, Python analysis of uploaded files, visual reasoning and image generation. OpenAI says o3 makes 20 percent fewer major errors than o1 on difficult, real-world tasks as judged by external experts, while o4-mini is a smaller, cost-efficient model that achieves 99.5% pass@1 on AIME 2025 with a Python interpreter. The company also introduced Codex CLI, an open-source coding agent for the terminal, and a $1 million grant initiative in $25,000 increments of API credits. OpenAI reports that both models remain below the "High" threshold of its Preparedness Framework in biological and chemical, cybersecurity, and AI self-improvement risk areas.

## Comments

**tim**: live stream

https://www.youtube.com/live/sq8GBPUb3rk?si=oWKIv3Akn44FP9qQ

**tim**: o3 is actually pretty nuts, the level of introspection and Googling that it does…

**tim**: Tyler Cowen says o3 is AGI

https://marginalrevolution.com/marginalrevolution/2025/04/o3-and-agi-is-april-16th-agi-day.html