OpenAI announced that starting from December 5, 2024, local time, there will be a 12-day (one session each working day) new product launch or demonstration event. Although the specific content of each launch cannot be determined yet, it may include the following based on various sources and speculations:

1. Public release of the text-to-video tool Sora

Sora is a new product that OpenAI publicly announced at the beginning of the year and has been in an invitation-only testing phase. The company’s former CTO Mira Murati has stated that Sora may be made publicly available to users by the end of the year, so the likelihood of Sora's public release during these 12 launch events is very high. In the field of artificial intelligence, text-to-video technology is a very hot research direction. For example, Google has also launched its latest generative AI video model, Veo. The release of Sora will provide users with new content creation methods, such as allowing users to input a story script, and Sora can generate corresponding video content, which has broad application prospects in film production, advertising creation, educational video production, etc.

2. A new reasoning model

According to reports from media outlets like The Verge, the new products include a new reasoning model, which may be a complete version of the o1 model or similar new reasoning AI models. Reasoning ability is crucial in artificial intelligence, such as in handling complex scientific and mathematical problems, and conducting logical analysis. OpenAI has previously planned to launch models focused on reasoning, such as the earlier mentioned 'Strawberry' model, which is also an AI product focused on reasoning. The new reasoning model may perform more efficiently and accurately in handling complex tasks, helping to enhance OpenAI's competitiveness in the field of AI reasoning, and may also be applied in scenarios that require complex logical reasoning, such as healthcare and scientific research.

3. New voice and functionality improvements in ChatGPT's voice mode

1. New voice

OpenAI may add a special voice to ChatGPT's voice mode (there are reports that it may be a Santa Claus voice), and users have already discovered that the voice mode button in the code can change to a snowflake shape, which may hint at the new voice. The addition of new voices can provide users with a more diverse interaction experience. For example, during specific holidays or themed scenarios, special voices can enhance fun and immersion.

2. Functionality improvements

There is room for improvement in the advanced voice mode of ChatGPT, which may implement real-time internet search and connect with cameras to analyze the real world. If real-time internet search is achieved, ChatGPT will be able to provide more timely and accurate information, no longer limited to the knowledge in the previous training data; connecting with cameras to analyze the real world can expand its applications in scenarios such as the Internet of Things and smart homes, such as recognizing indoor environmental conditions and conducting security monitoring.

4. Other potential release contents

1. Unlocking the image generation capabilities of the GPT - 4o model

The image generation capabilities of the GPT - 4o model, which have been locked, once showed impressive effects and generation control ability in early demonstrations; this model also has native video analysis capabilities. If the image generation feature is unlocked, GPT - 4o will have more comprehensive capabilities in multimedia content generation, allowing users to utilize it for image creation, design, and other tasks.

2. Preview of the o2 model or GPT - 5o

Netizens are very much looking forward to the preview of the o2 model or GPT - 5o. If relevant previews can be seen at the launch event, it will allow users and developers to gain insight into the future development direction of OpenAI's models, potentially involving information on architecture, performance enhancements, new features, etc., which may also guide industry development trends.

3. Early reveal of the text-to-speech tool and the AI agent framework codenamed 'Operator'

OpenAI's text-to-speech tool and the AI agent framework codenamed 'Operator' are expected to be released next year, but there is also a possibility they may debut within the next two weeks. The text-to-speech tool can play a role in voice interaction, audiobook production, and other fields; the AI agent framework helps build smarter and more efficient AI agents, applicable in intelligent customer service, automated process management, and more.

$IOTX