TechFlow news, on December 6, OpenAI announced the full rollout of its upgraded o1 model in ChatGPT, replacing the previous o1-preview version. According to the company, the new version achieves significant breakthroughs across several key areas: faster reasoning speed, improved performance in programming, mathematics, and writing, new image upload and visual analysis capabilities, and a 34% reduction in error rates when handling complex problems.
On the business model front, OpenAI has introduced a tiered subscription strategy: ChatGPT Plus users ($20/month) can immediately access the o1 model, while a newly added Pro tier ($200/month) offers enhanced reasoning capacity and unlimited access. Enterprise and education plan users will gain access within one week.
Regarding the product roadmap, OpenAI plans to release an API for the o1 model, supporting function calling, developer messages, structured outputs, and visual capabilities. Additional tool features such as web browsing and file uploads are also under development. The company has also published an updated system safety evaluation report, which, based on an expanded test suite, maintains the o1 model's risk rating at "medium," indicating it is safe for deployment.
