Google unveils its latest AI model to compete with DeepSeek
What's the story
Google has unveiled its latest artificial intelligence (AI) model, the Gemini 2.0 Pro Experimental. The announcement comes alongside other AI model releases, including the 'reasoning' model called Gemini 2.0 Flash Thinking, which is now available in the Gemini app. The developments come at a time when the tech industry is increasingly interested in cost-effective AI reasoning models like those offered by Chinese firm DeepSeek.
Market competition
New AI model to compete with DeepSeek's R1
Both Google and DeepSeek launched their AI reasoning models in December. However, DeepSeek's R1 got more attention owing to its better performance and affordability. In a bid to counter that, Google is now promoting its Gemini 2.0 Flash Thinking model via the popular Gemini app. The tech giant accidentally announced the release of Gemini 2.0 Pro a week ago in the app's changelog but has now officially confirmed its launch.
Model capabilities
Gemini 2.0 Pro: Google's most advanced AI model yet
The Gemini 2.0 Pro, an upgrade over last year's Gemini 1.5 Pro, is now the flagship model of Google's Gemini AI family. It will be available on AI development platforms Vertex AI and Google AI Studio, as well as for Gemini Advanced subscribers in the app. The new model is better at coding and complex prompts with higher world knowledge understanding than its predecessors, Google said.
Processing power
It can process 1.5 million words at once
The Gemini 2.0 Pro features a context window of two million tokens, which means it can process some 1.5 million words at once. To put this into perspective, it could ingest all seven books in the Harry Potter series in a single prompt and still have some 400,000 words to spare. This processing power is a major highlight of Google's latest AI model.
New releases
Gemini 2.0 Flash and a cost-efficient model
Along with the Gemini 2.0 Pro, Google is also rolling out its Gemini 2.0 Flash model to all Gemini app users on both desktop and mobile. In an effort to compete with DeepSeek's models, Google is also introducing a new and more cost-efficient AI model, Gemini 2.0 Flash-Lite. The company claims that the model outperforms its predecessor but works at the same price and speed as the previous version.
Information
Google's AI investment set to more than double in 2025
Google plans to invest a whopping $75 billion into its AI models this year, a sharp increase from the $32.3 billion in 2023. The increase in investment comes as Google tries to keep up with other AI players like OpenAI, Microsoft, and Amazon-backed Anthropic.