12 Days of OpenAI: A Roundup of Every Day’s Drops
What are the ’12 days of OpenAI’?
OpenAI announced a 12-day event series, starting on December 5, featuring live streams and the release of "a bunch of new things, big and small." The series concluded on December 20, with the biggest announcement of the year.
What’s dropped?
Friday, December 20
- OpenAI unveiled its latest models, o3, which encompass o3 and o3 mini.
- o3 can outperform o1 in various benchmarks, including math and science.
- o3 mini is a new model in the o3 family that supports three reasoning options: low, medium, and high.
- OpenAI is opening up the o3 models to external safety testing.
- The o3 model is planned to launch at the end of January, with the full o3 model launching after that.
- The company also introduced deliberative alignment, a training paradigm that teaches reasoning LLMs the text of human-written and interpretable safety specifications.
Thursday, December 19
- OpenAI released updates regarding its MacOS desktop app and its interoperability with other apps.
- Users can now use the desktop app on MacOS to see and automate their work with ChatGPT.
- The app now supports Apple Notes, Quip, and Notion for writing.
- The app also supports Advanced Voice Mode while working with other apps.
Wednesday, December 18
- OpenAI released a toll-free phone number, 1-800-ChatGPT, allowing users to access ChatGPT without Wi-Fi.
- Users can call the number to access ChatGPT from anywhere in the US, and from other countries through WhatsApp.
Tuesday, December 17
- OpenAI released developer features and updates, dubbed "Mini Dev Day."
- The o1 model is now out of preview in the API, with support for function calling, structured outputs, developer messages, vision capabilities, and lower latency.
- The Realtime API now supports WebRTC, and has a 60% audio token price drop.
- The fine-tuning API now supports Preference Fine-Tuning.
- OpenAI introduced new Go and Java SDKs in beta.
Monday, December 16
- OpenAI released Search in ChatGPT, available to all users.
- The AI search engine allows users to search the web, translate text, and more.
- The feature is available on all devices, and users can access it by clicking on the magnifying glass icon.
Wednesday, December 11
- Apple released iOS 18.2, which includes integrations with ChatGPT across Siri, Writing Tools, and Visual Intelligence.
- Siri can now recognize when a user asks a question outside its scope that could benefit from being answered by ChatGPT.
- Visual Intelligence allows users to point their camera at something and search the web with Google, or use ChatGPT to learn more about what they are viewing.
Tuesday, December 10
- OpenAI released Canvas, a new feature that allows users to create and edit text-based documents.
- Canvas is now available to all web users, regardless of plan.
- Canvas has been built into GPT-4o natively, allowing users to create and edit documents directly in the model.
Monday, December 9
- OpenAI released Sora, a new video model that can generate video-to-video, text-to-video, and more.
- Sora is available to ChatGPT Plus and Pro users.
- The model can generate up to 50 videos per month at 480p resolution or fewer videos at 720p.
Friday, December 6
- OpenAI expanded access to its Reinforcement Fine-Tuning Research Program.
- The program allows developers and machine learning engineers to fine-tune OpenAI models to excel at specific sets of complex, domain-specific tasks.
- OpenAI encourages research institutes, universities, and enterprises to apply to the program.
Thursday, December 5
- OpenAI started the event series with the release of ChatGPT Pro and the full version of the o1 model.
- ChatGPT Pro is a new tier of subscription that grants unlimited access to the best OpenAI has to offer.
- The full version of o1 is available to all ChatGPT Plus and Pro users, and features multi-modal reasoning.
Where can you access the live stream?
The live streams were held on the OpenAI website, and posted to its YouTube channel immediately after.
