The next step for OpenAI’s logic models is o3, a design previewed on Dec. 20. o3 and its smaller cousin, o3-mini, outperformed o1 in programming, math, science, and ‘ philosophical logic’ tests designed to evaluate human-like knowledge and research applications.  ,’ Reasoning’ includes a security feature called democratic position, in which the unit uses a” chain of thought” to stop users from jailbreaking or tricking it into bypassing safety measures.
However, Google’s Gemini 2.0 Flash Thinking Experimental design treads equivalent ground to OpenAI o1’s logic abilities.
‘ 12 Times of OpenAI ‘ brings new equipment and new conceptual AI features
The o3 news came at the close of OpenAI’s” 12 Days of OpenAI” plan, a holiday time line of product changes. These announcements, from Dec. 5 to Dec 20 ( excluding weekends ), showcased new features for OpenAI’s generative AI tools, with some available now and others still in testing.
Day 1: The$ 200 ChatGPT Pro and o1 changes
On Dec. 5, OpenAI introduced a new subscriber level for ChatGPT: the Pro program. For$ 200 per month, the Pro subscription brings OpenAI o1, o1-mini, GPT-4o, and Advanced Voice to ChatGPT. Additionally, it gives access to o1 professional mode, a more computation-intensive version made for challenging issues that professional engineers and researchers must overcome.
On the same day, OpenAI announced an updated, more comprehensive system card for the hotly-anticipated o1 design.
Day 2: The Reinforcement Fine-Tuning Research Program
OpenAI developed a new tool to allow machine learning engineers and developers to create custom designs for certain things with the Reinforcement Fine-Tuning Research Program. Quick 2025, it is anticipated to go public with beta testing.
Time 3: Sora film engine
OpenAI’s realistic picture engine, announced early next year, is now available for ChatGPT Pro customers. While Artificial video design is easier than ever, types like Sora also struggle with complex, fast-moving content and can often be identified by a too-perfect shininess. YouTube videos will be watermarked in accordance with C2PA guidelines to indicate that they are AI-generated.
SEE: Using some of the numerous free Microsoft and LinkedIn courses that have been updated for 2024, you can learn the fundamentals of generative AI.
Day 4: Canvas
Canvas, a coding interface introduced in beta in October, became generally available in December. Developers can connect to their apps using the current version of Canvas because it comprehends and writes Python and integrates with custom GPTs. It also allows users to view prompts and outputs side-by-side for easier reference.
Day 5: Apple on-device AI with ChatGPT
During the 12 days of OpenAI, Apple Intelligence received its anticipated ChatGPT update. For more complex queries that the on-device chip cannot handle, the on-device Apple Intelligence can now access ChatGPT servers.
Day 6: Advanced Voice with Video
Advanced Voice mode, available to ChatGPT subscribers, can now converse about images on your computer screen or through your camera. The audio version of the chatbot receives more flexible responses and natural speech thanks to the mode.
Day 7: Projects
As of Dec. 13, ChatGPT Plus, Pro, and Team users can organize their chats into Projects, or separate instances. Projects allow users to set up specific instructions that only apply to Projects, and relevant resources can be backed up with it. In January, users of Enterprise and Edu will have access to this feature.
Day 8: ChatGPT search upgrades
After the December release, ChatGPT search underwent several changes, including a new map interface, faster mobile response times, and more Advanced Voice functionality to keep it up to par with the rest of the paid-tier voice offerings. Users can now access search at the free tier as long as they log in with an email address.
Day 9: New features, options, and upgrades for developers
Day nine was all about developers, with a variety of announcements:
- Developers can now use the API to access OpenAI o1.
- Various upgrades for the API were released, including a simpler WebRTC integration, a 60 % price reduction for GPT-4o audio, and support for GPT-4o mini at one-tenth of previous audio rates.
- Preference fine-tuning allows for improved customization.
- The Java SDKs for Go and Java are now available for beta testing.
Day 10: 1-800-CHATGPT
OpenAI’s generative AI has launched a phone and WhatsApp line, akin to Google’s classic voice search. Users can ask natural-language questions, and the chatbot will respond for free. OpenAI considers this feature experimental, noting that its availability and limitations may change.
Day 11: More options for apps
Day 11 brought a long list of connections from ChatGPT to more coding apps and tools, including VS Code forks, Jetbrains IDEs, additional Terminal apps, and more. ( Initially, it supported iTerm 2, Terminal, TextEdit, VS Code, and Xcode. )  , Three new app integrations arrived, connecting ChatGPT to Apple Notes, Notion, and Quip. Advanced Voice Mode now supports a variety of other user-defined desktop apps.
OpenAI notes that ChatGPT won’t interact with desktop apps without the user’s permission.
Plus, Pro, Team, Enterprise, and Edu users can use the new app integrations.
Day 12: o3 and o3-mini
OpenAI saved the biggest news for last: o1 is no longer the company’s foremost model. Instead, o3 – now in early access for safety and security researchers – improves coding, math, and science performance. The company also pioneered a new technique called deliberative alignment, used to keep o3 on-mission. Researchers working in safety can apply to test o3 here.
