If you use Gemini as your virtual assistant, we have good news for you. Google has just updated the mobile and desktop apps to include access to two of its most powerful AI models: Gemini 2.0 Flash Thinking and an experimental version of Gemini 2.0 Pro.
Please follow us on Facebook and Twitter.
Gemini 2.0 Flash Thinking Experimental was launched in December 2024, but until now, it has only been available through the Gemini API and Google AI Studio. Starting now, it is also part of the AI chatbot apps, and it’s available for free.
Let’s remember that this AI stands out for its “reasoning” capabilities, similar to those used by OpenAI in the o1 and o3 models, or DeepSeek in R1. This means users can see how the technology analyzes and “thinks” about their answers before including them in its final response.
Google also introduced another version of Gemini 2.0 Flash Thinking that can interact with other apps and services in its catalog, such as Google Maps, YouTube, and its web search engine. Google claims that this AI is the best in the world today, as it currently holds the top spot in the Chatbot Arena LLM ranking.
Google has launched an experimental version of Gemini 2.0 Pro.
The company has also integrated this version into the Gemini app. Google claims this technology excels in programming tasks and handling complex prompts. It also outperforms other models in areas such as reasoning, mathematics, language translation, and video analysis, among others.
Gemini 2.0 Pro also includes a 2 million token context window, the largest Google has ever implemented. However, to access Gemini 2.0 Pro in the Gemini apps, you need to subscribe to Gemini Advanced, which costs $20 per month. You can try it for free for one month.
Those interested in trying Gemini 2.0 Flash Thinking and the experimental version of Gemini 2.0 Pro can do so from the model selection menu in the mobile and desktop apps.
Gemini is evolving rapidly.
Google also announced that Gemini 2.0 Flash is now available to developers through Vertex AI and Google AI Studio. Additionally, the company launched a new variant called Gemini 2.0 Flash-Lite, which is now available for testing as a public preview on its developer platforms.
Mountain View reports that Flash-Lite has a 1 million token context window and supports multimodal input methods. In terms of performance, Google says Flash-Lite outperforms Gemini 1.5 Flash, surpassing it in most benchmarks while maintaining the same speed and price.
