Google Announces Thinking Model with Gemini, following OpenAI and DeepSeek

Google has released its latest Gemini 2.0 Flash Thinking Experimental model, along with a slew of other AI models, including the Gemini 2.0 Pro Experimental. The new flash thinking model, which was released for Gemini app users, now adds thinking and reasoning capabilities to Gemini, similar to what OpenAI has done with the o1 and DeepSeek's R1 models.

Notably, Google released its reasoning model in December but received little traction at the time. The world is now fixated on reasoning models, due to DeepSeek, which appears to have prompted Google to bundle its model with the Gemini app. In a blog post announcing the launch, Google stated that the Gemini 2.0 Flash model is "currently ranked as the world's best model."

Just days ago, during the Q4 earnings call, Google announced that it expects to spend $75 billion on AI infrastructure as it competes with OpenAI, Microsoft, and Meta, as well as a growing presence of low-cost Chinese models.

Google is also releasing a version of 2.0 Flash Thinking that is compatible with apps such as YouTube, Search, and Google Maps. Google's thinking model works exactly as you'd expect: it displays its thought process, allowing you to see why it responds to queries and prompts in a specific way, as well as trace the model's reasoning.