OpenAI has launched o1, a new family of AI models to fact-check and generate code.

o1 preview uses a "chain of reasoning" approach, considering all parts of a question before responding. Additionally, the model can plan and perform a series of actions over an extended period to arrive at an answer, making it suitable for tasks requiring the synthesis of multiple subtasks. Furthermore, the pricing for o1-preview in the API is USD 15 per 1 million input tokens and USD 60 per 1 million output tokens.

o1-mini is a cheaper reasoning model for coding and is 80% cheaper than o1-preview.

Analyst QuickTake: This news comes as the company reportedly seeks to raise USD 6.5 billion in funding while considering restructuring to attract investors. Notably, the new o1 model was code-named "Strawberry" and planned to be launched in the fall of 2024 (September–December). The company claims that the model outperforms previous data analysis, science, and coding models. However, o1 can be slower than other models, sometimes taking over 10 seconds to answer questions, and may still produce errors and hallucinations.