OpenAI Tests New AI Models: o3 and o3 Mini

(Reuters) – OpenAI announced on Friday that it is testing new reasoning AI models, o3 and o3 mini, indicating increased competition with rivals such as Google in the pursuit of more advanced models capable of addressing complex problems.

Upcoming Launches

CEO Sam Altman stated that OpenAI intends to launch o3 mini by the end of January, followed by the full release of o3. These more robust large language models are anticipated to outperform existing models, potentially attracting new investments and users.

Previous Releases

In September, Microsoft-backed OpenAI released the o1 AI models, which are designed to spend more time on processing queries to tackle difficult problems. According to a blog post, these o1 models can reason through complex tasks and solve more challenging issues in science, coding, and math.

Enhanced Capabilities

The new o3 and o3 mini models, currently undergoing internal safety testing, are expected to be more powerful than the previously launched o1 models, according to the company.

Research Opportunities

OpenAI has announced an application process for external researchers to test the o3 models ahead of their public release, which will close on Jan. 10.

Industry Impact

Since launching ChatGPT in November 2022, OpenAI has ignited an AI arms race. The company's rising popularity and new product launches contributed to a successful $6.6 billion funding round in October. Rivals like Alphabet's (NASDAQ: GOOGL) Google have also stepped up, having released the second generation of their Gemini AI model earlier in December, as they aim to regain the lead in the AI technology race.