San Francisco, October 29, 2024 – OpenAI is primed to launch what it boasts is an advanced version of ChatGPT, named Orion, before the end of December 2024, and this release would constitute a major development in the abilities of artificial intelligence systems. The company, however, intends to confine the first batch of its model to a few partner companies that have been given access.
“We don’t have plans to release a model code-named Orion this year,” stated Niko Felix, OpenAI spokesperson, while noting that “we do plan to release a lot of other great technology.”
Despite these official statements, internal sources suggest the development of Orion is at its final stages.
The new model reportedly demonstrates computing power that is up to 100x greater than its predecessor, GPT-4. Microsoft, which is the main partner of OpenAI, is reportedly readying its Azure cloud service for the launch of Orion, with sources indicating potential operational readiness by November.
There have been reports that OpenAI leveraged its internal o1 model ‘Strawberry’ to build synthetic datasets for the training of Orion itself. The latter has been making rounds after a post by Chief Executive Officer Sam Altman referencing ‘winter constellations,’ which many interpreted as the working title of this project.
The release of Orion comes amidst other changes – the restructuring that followed the exit of CTO Mira Murati and other senior officers barely a month after a $6.6 billion round of investment that evidently changed the company’s profile.
There are speculative reports that the training phase of the Orion project was completed by OpenAI at the end of September, though the company maintains its characteristic secrecy regarding specific capabilities and release timelines.
Latest Stories:
Decagon Achieves 91% Customer Support Automation Using OpenAI Models
AI Infrastructure Growth Threatens Global E-Waste Crisis, Research Warns
Central Wisconsin Launches Revolutionary AI Center for Business and Workforce Training