12 Days of OpenAI Recap

Day 1: The full version of the o1 model was launched, offering significant improvements in speed, intelligence, and multimodality. Additionally, ChatGPT Pro was introduced, featuring o1 Pro mode for tackling complex tasks.

Day 2: OpenAI previewed reinforcement fine-tuning (RFT), which allows users to customize models with their own data using reinforcement learning.

Day 3: Sora, a new video generation model, was launched and is available to ChatGPT Plus and Pro users.

Day 4: Canvas was released to all users, enabling collaborative writing and coding with ChatGPT, including the ability to run Python code.

Day 5: ChatGPT was integrated with Apple Intelligence, allowing users to access ChatGPT directly through their Apple devices using Siri, writing tools, and camera control.

Day 6: Video and screen sharing were added to Advanced Voice mode, along with the launch of Santa Mode.

Day 7: Projects were launched, enabling users to organize conversations, upload files, and set custom instructions within ChatGPT.

Day 8: Search was rolled out to all logged-in free users, along with integration of search into advanced voice mode.

Day 9: A mini Dev Day focused on developers and startups, with the release of o1 in the API, along with function calling, structured outputs, developer messages, vision inputs, and a reasoning effort parameter, and other updates including the realtime API updates, and preference fine-tuning.

Day 10: OpenAI introduced 1-800-CHAT-GPT, enabling users to interact with ChatGPT via phone calls or through WhatsApp messaging.

Day 11: ChatGPT gained native app integrations and supporting tools like Quip, VS Code, and Apple Notes for seamless workflow integration.

Day 12: OpenAI announced the release of the new o3 and o3-mini reasoning models, representing the next generation of advanced AI capabilities.

Check out the full videos here.