In December 2023, Google introduced Gemini, its new AI model, and is planning to launch Gemini 2.0 in December 2024 (reported by The Verge). The information states Google aims to make this new version widely available at that time. However, it seems the new model isn’t performing as well as the team, led by Demis Hassabis, had hoped. This isn’t just a Google issue; many companies developing large AI models are facing similar challenges.
Back in December, Google shared details about Gemini 1.0 and released Gemini 1.0 Pro through Bard, which later got a name change. They also released a version called Nano for the Pixel 8 Pro.
By February, a more advanced version, Gemini 1.0 Ultra, was available for those subscribed to Gemini Advanced, and it offered an expanded context window, which means it could understand and process more information at once. Subscribers had access to this new feature starting in May.
When Google talks about a “widely release,” it likely means they want to make the new model available to both developers and regular users, similar to how they did with earlier versions.
At the same time, people are still waiting for Project Astra, which is supposed to add camera and vision capabilities to Gemini Live. It’s not clear how Google will handle both announcements since they might happen around the same time.
Interestingly, OpenAI is also rumored to be releasing its next big model in December. However, they have denied reports about it. The new model, called Orion, is expected to have a limited launch first with selected partners before being available to everyone.
If everything goes according to plan, Google will have a yearly release schedule for its major models, aligning with their developer conference in May. This regular schedule keeps users excited and engaged with the latest AI advancements.