Google Unveils Gemini 2.0 Flash AI Reasoning Model To Challenge OpenAI's o1

OpenAI has been in the headlines as the company has been introducing several new products alongside a new pricing tier, but now Google is trying to make some noise of its own. The company’s newest model, which is referred to as Gemini-Exp-1206, is based on Flash 2.0, which is a complex reasoning model meant to tackle math, coding, and instruction following tasks. It’s meant to be a direct competitor to OpenAI o1 model.

Gemini-Exp-1206 is an experimental model and will be available to users who are Gemini Advanced subscribers. To access this new model, users will need to use the drop-down menu available on the web and desktop app. It will appear to users as “2.0 Experimental Advanced” while “Preview gemini-Exp-1206” will show up underneath it, which is a cumbersome implementation. Google could have done a better job naming this model to make it easier to users to understand what they’re using.


Google is noting that because this is an experimental model, that there will be several limitations and that users should expect to see unexpected behavior. Other drawbacks include a lack of real time information and limited functionality with Gemini features users have become accustomed to. Despite these rough edges, Google claims that the model will deliver significant performance improvements and is ready to help users with school projects or develop business plans.

The naming schemes and listed limitations make this effort feel a little rushed. As if Google felt it needed to get its most cutting-edge model out the door, warts and all, to show that it can keep up with OpenAI. Hopefully the company can iterate quickly on its path to making it a fully baked product, so that paying Gemini Advanced customers are getting good value from the service.