OpenAI Unveils "o1" Reasoning Model, Kicking Off "Shipmas" on December 5, 2024

OpenAI officially launched its highly anticipated "o1" reasoning model to ChatGPT users on December 5, 2024, marking the commencement of the company's "12 Days of OpenAI" event, colloquially known as "Shipmas." This release introduced a new class of AI designed to enhance complex problem-solving and logical deduction capabilities, moving beyond traditional generative pre-trained transformer (GPT) numbering. The full version of o1 followed a preview release of o1-preview and o1-mini earlier in September 2024.

The launch was accompanied by a playful social media remark from a user identified as "Lisan al Gaib," who stated, "yes I pulled the naming and the date out of a hat." The user further elaborated on the naming choice, noting, "I feel like GPT-5.2 wouldn't really do the model justice," and confirmed the launch date, "and Dec 5th, because that's when o1 was launched on the first day of shipmas last year." This suggests a deliberate move by OpenAI to distinguish o1 as a unique advancement.

OpenAI's o1 model is engineered to spend additional time "thinking" before generating responses, significantly improving its performance on complex reasoning tasks, scientific problems, and programming challenges compared to its predecessor, GPT-4o. This approach has led to a reported 34% reduction in major errors on difficult problems. The model was initially made available to ChatGPT Plus and Team subscribers, with API access for developers in higher usage tiers.

Alongside the o1 launch, OpenAI introduced ChatGPT Pro, a new premium subscription tier priced at $200 per month. This Pro tier offers unlimited access to o1, GPT-4o, and an exclusive "o1 pro mode," which utilizes more computational power for even more refined answers. The "12 Days of OpenAI" event, which began on December 5, 2024, featured a series of daily updates and product releases, highlighting OpenAI's rapid innovation pace.