What you need to know
- Google is reportedly gearing up to launch Gemini 2.0, its next artificial intelligence model.
- Now, a new model titled Gemini-2.0-Pro-Exp-0111 has appeared for at least one user of the Gemini web client.
- It’s unclear if this curiously-named AI model is truly powered by Gemini 2.0 Pro, or when the model will become widely available.
Google and OpenAI are both planning to release the next versions of their flagship AI models by year’s end, or in early 2024. Prior reports indicated that the two companies are gearing up to ship Gemini 2.0 and Orion, respectively, sometime in December. More recently, the folks at TestingCatalog spotted a new and conspicuously-named Gemini model that could be Gemini 2.0 Pro.
TestingCatalog says that the model is “quite fast” responding to prompts, but that it’s impossible to know whether it’s really Gemini 2.0 Pro under the hood. It’s named Gemini-2.0-Pro-Exp-0111, and was discovered in the dropdown menu that lets you select which model the Gemini chatbot uses. Currently, the option for free users is Gemini 1.5 Flash. Paid subscribers gain access to Gemini Advanced, which adds Gemini 1.5 Pro.
The new Gemini-2.0-Pro-Exp-0111 model appears under the Gemini Advanced section in the dropdown menu, signaling that it will begin as a paid exclusive. However, it’s possible (and perhaps likely) that Google is working on a Gemini 2.0 Flash model that could replace Gemini 1.5 Flash for free users.
I can already prompt it and get a response. It seems to be working quite fast but I am still uncertain if the response comes from 2.0 for real 🤖
It also can draw images and use web search! pic.twitter.com/PkYgGCcvVpNovember 7, 2024
There is reason to believe that Gemini-2.0-Pro-Exp-0111 is really Gemini 2.0 Pro, like how it’s labeled as “our experimental model” by Google. However, there’s no way to know for sure. It could be a user-interface test, or simply appear by accident. Being that prior reports expected a December 2024 release for Gemini 2.0, it’s reasonable that Google would be expanding tests for the model in the days or weeks immediately preceding its launch.
TestingCatalog confirms that Gemini-2.0-Pro-Exp-0111 can respond to prompts, draw images, and search the web. It’s not available to the public yet, so you won’t find it by going to gemini.google.com. A release sometime next month would mark exactly one year after the first Gemini model debuted. As a reminder, Gemini now references the underlying model (like the upcoming Gemini Pro 2.0 model), and the interface that lets you interact with the model (such as the Gemini web client).
By all accounts, the Gemini 2.0 Pro model will be made available as part of the Google One AI Premium subscription, which costs $20 per month and includes Gemini Advanced.