OpenAI, the company behind ChatGPT, has launched an exciting series of pre-holiday events, revealing new AI features and products daily for 12 consecutive days. Starting on December 5 at 10 a.m. Pacific Time (8 p.m. Kyiv time), the event promises a mix of major releases and smaller, intriguing updates. As stated by OpenAI CEO Sam Altman on the X platform (formerly Twitter), “Every day we will be live-streaming new products or demos — including both large-scale launches and nice little things. We have a lot to show, we hope you enjoy! Merry Christmas.”
Though OpenAI has been tight-lipped about specific details, the AI community eagerly anticipates new releases, including the much-anticipated text-to-video tool, Sora, and an improved model in the o1 series designed for enhanced logical reasoning. Among the more subtle updates, users can expect a new Santa-themed ChatGPT voice and a festive snowflake icon.
Day One: Major Updates and New Plans
OpenAI kicked off its holiday series with a major update to the o1 logical thinking model and the introduction of a new ChatGPT Pro tariff plan. The o1 model, codenamed Strawberry, has now moved beyond pre-access and is available to users on the ChatGPT Plus and Team tariffs. From next week, it will also be available to Enterprise and Edu customers. This updated version is faster, more powerful, and more accurate than its predecessor, o1-preview, offering notable improvements in coding, mathematics, and image processing capabilities.
In addition to the o1 model update, OpenAI announced a new ChatGPT Pro plan priced at $200 per month. This plan provides unlimited access to o1, GPT-4o, and the Advanced Voice mode. It also features a Pro mode for o1, offering enhanced computing resources to solve complex problems. The previous Plus plan remains at $20 per month, though it will rise to $24 in 2025 for Ukrainian users. The Plus plan grants early access to new features and all models except the more advanced o1 version.
As part of its holiday campaign, OpenAI also unveiled the ChatGPT Pro Grant Program, which aims to support medical research, with plans to expand this initiative into other industries.
Day Two: Expansion of Reinforcement Fine-Tuning
On the second day, OpenAI introduced the expansion of its reinforcement fine-tuning research program. This initiative enables developers and researchers to fine-tune AI models for specialized tasks within niche industries, boosting accuracy and performance.
Reinforcement fine-tuning is a method where developers influence the behavior of AI models by providing specific tasks and feedback, which helps improve the model’s accuracy in solving similar problems in the future. For example, by training a model with legal text and evaluating its responses, developers can enhance the model’s understanding of legal terminology.
This approach holds significant potential for AI applications in specialized fields such as medicine, finance, and scientific research. OpenAI is opening this program to research institutes, universities, and enterprises that work on well-defined tasks that can benefit from AI optimization.
Day Three: The Long-Awaited Sora Model
Day Three was marked by the eagerly awaited announcement of Sora, OpenAI’s new AI model capable of converting text into video. OpenAI first demonstrated Sora in February 2024, but now the model is being made available to a broader audience.
The core functionality of Sora — generating videos from simple text prompts — remains unchanged. However, OpenAI has significantly enhanced the speed and reduced the cost of using the model. Several new features have been introduced, including:
- Storyboard: This feature allows users to create multiple videos and combine them into a single clip using an interactive timeline, similar to professional video editing software like Adobe Premiere Pro.
- Creativity Feed: A platform where users can publish their videos and explore others’ creations, modifying them as needed. Users can also view the queries that generated each video.
Sora also includes the Remix tool, which allows users to change the video style, and offers a range of ready-made styles. Users can trim, loop, and choose from three video formats: vertical, horizontal, and square.
OpenAI’s paid subscribers can access Sora’s video generation features:
- ChatGPT Plus ($20/month): Create up to 50 videos per month, each up to 5 seconds long and at 720p resolution.
- ChatGPT Pro ($200/month): Create longer videos (up to 20 seconds), with higher resolution (1080p), and generate up to five variations of a single query. The limit is 500 videos per month.
Sora operates as a standalone product. To access it, users need to log in with their ChatGPT Plus or Pro account on sora.com.
Day Four: Canvas Update and Accessibility for All Users
On the fourth day, OpenAI presented an update to Canvas, a feature introduced earlier in October. Canvas was originally designed for programming and writing tasks and is now available to all users, including those on the free tier. It is integrated directly into ChatGPT, opening in a separate window where ChatGPT acts as a text editor or code reviewer.
Canvas is designed to help users interact with AI in the context of a project, making it easier to get recommendations and suggestions. The interface functions like a shared document, but instead of collaborating with people, users work with ChatGPT, which provides insights directly within the project.
This feature is now accessible via the web version and the ChatGPT Windows app, with macOS support expected soon, notes AIN.
Day Five: ChatGPT Integration with Apple
On Day Five, OpenAI introduced ChatGPT’s integration with Apple Intelligence for iOS and macOS, a feature that Apple has begun rolling out to its devices. During the demonstration, OpenAI’s team showcased how Siri can provide responses generated by ChatGPT, along with links to view the full answers in the ChatGPT app.
The integration focuses on three main areas:
- Siri Voice Commands: Using ChatGPT to enhance Siri’s responses.
- Writing Tools: For editing and improving text.
- Visual Intelligence: With a new camera button on the iPhone 16, users can interact with ChatGPT’s visual features.
While basic features are available without an account, advanced versions of ChatGPT can be accessed through paid Apple subscriptions.
Day Six: Multimodal Voice Features
On Day Six, OpenAI unveiled an enhanced voice mode with video capabilities. ChatGPT now can watch and interact with users in real time during conversations, making it feel more like talking to a human. Additionally, a Santa-themed voice, complete with a British accent, was added for the holiday season.
The new voice mode is multimodal, supporting video, speech, and text, making interactions feel more natural. This mode also includes a screen-sharing feature, allowing users to share their apps with ChatGPT for help. Paid subscribers to ChatGPT Plus and Pro can already access this feature, with enterprise and education customers receiving access early next year.
Day Seven: New Project Organization Tool
On Day Seven, OpenAI introduced Projects, a new feature designed to help users organize and manage their chats. This update brings together various ChatGPT features in one place, including Canvas. Projects are aimed at helping users stay organized and better manage multiple tasks simultaneously.
During the demo, OpenAI showed how Projects could be used for fun activities like organizing a Secret Santa exchange. Users can upload wish lists, assign people to each other, and have ChatGPT draft the necessary emails. The feature is available to ChatGPT Plus, Pro, and Team users.
Day Eight: ChatGPT Search for All Users
Day Eight brought a significant update with the global rollout of ChatGPT Search. Initially available only to paid users, this feature is now accessible to all registered users worldwide. ChatGPT Search allows users to search the web without leaving the chat interface. Feedback from users has led to new improvements, including the ability to search during voice conversations.
By enabling a special button, users can have ChatGPT perform web searches by default, providing results similar to Google, with links for further information.
Day Nine: Developer Tools and New APIs
On Day Nine, OpenAI rolled out its enhanced o1 model to developers via the API. This model, the successor to o1-preview, offers several improvements, including support for API calls, structured responses, and image processing. It is designed to help developers build applications for tasks like voice assistants and virtual tutors.
Additionally, OpenAI introduced official SDKs for Go and Java, making it easier for developers to integrate OpenAI models into their applications, regardless of the programming language they use.
Day Ten: ChatGPT via Phone
On Day Ten, OpenAI expanded ChatGPT’s accessibility by making it available via phone, even on landlines in the US. Users can now call 1-800-242-8478 for 15 free minutes of AI interaction per month. People in other countries can access ChatGPT via WhatsApp, though voice conversations and photo analysis are not yet available in this format.
Day Eleven: Expanded MacOS Support
On Day Eleven, OpenAI expanded support for ChatGPT on macOS, adding compatibility with a range of third-party coding and productivity apps. This includes support for popular coding environments like BBEdit, MatLab, and IntelliJ IDEA, as well as applications like Notion and Apple Notes.
Soon, Windows users will also have access to these features, and the company promises additional improvements in the near future, adds NIXsolutions.
Day Twelve: A Special Surprise
OpenAI teased that the 12th day would bring something “super special,” though the details remain under wraps for now. As always, we’ll keep you updated on any developments as they unfold.