- OpenAI launched a marketing campaign called “Shipmas”.
- The campaign is expected to include 12 days of product releases, demos and new features.
- On the first day, OpenAI released an updated version of o1 and ChatGPT Pro.
OpenAI releases new features and products before the holidays.
Fans are hoping the 12-day campaign, which OpenAI calls “Shipmas,” will include an update to its text-to-video AI tool, Soraas well as other updates to its ChatGPT and o1 models.
OpenAI hinted at its plans Wednesday in a post on
CEO Sam Altman also hinted at the campaign Wednesday during the New York Times’ DealBook Summit. “We have a bunch of cool new stuff. We’re doing this kind of fun holiday stuff. We’re doing 12 days of OpenAI,” he said. “We’ll either launch something or we’ll do a demo every day for the next 12 days of the week.”
Here’s everything OpenAI has released so far for “Shipmas.”
‘Shipmas’ Day 1
OpenAI kicked off the promotion with a bang by releasing the full version of his last model of reasoning, o1.
OpenAI preview o1 in September, describing it as a series of artificial intelligence models “designed to spend more time thinking before responding.” Until now, only a limited version of these templates was available to ChatGPT Plus and Team users.
These users now have access to the full capabilities of the o1 models, which Altman says are faster, smarter and easier to use than Preview. They are also multimodal, meaning they can process images and text jointly.
Max Schwarzer, a researcher at OpenAI, said the full version of o1 was updated based on user feedback from the pre-release version and said it was now more intelligent and accurate.
“We performed a fairly detailed suite of human evaluations for this model, and found that it made major errors about 34% less often than the o1 preview, while thinking fully about 50% faster,” said he declared.
With o1, OpenAI has unveiled a new level of ChatGPT called ChatGPT Pro. It is priced at $200 per month and includes unlimited access to the latest version of o1.
Day 2 of “Shipmas”
On Friday, OpenAI previewed an advancement that allows users to fine-tune o1 on their own datasets. Users can now leverage OpenAI’s reinforcement learning algorithms, which mimic the human learning process through trial and error, to customize their own models.
The technology will be publicly available next year, allowing anyone from machine learning engineers to genetic researchers to create domain-specific AI models. OpenAI has already partnered with Reuters news agency to develop a legal assistant based on o1-mini. He also partnered with the Lawrence Berkeley National Laboratory to develop computational methods to assess rare genetic diseases.