The technology giant, Google, continues to push the boundaries of artificial intelligence with the launch of Gemini 3 Deep Think. Designed for Google AI Ultra subscribers, this model promises to transform the way scientists and developers approach complex problems. Discover the impressive capabilities of this advanced reasoning system.
The 3 must-know facts
- Gemini 3 Deep Think, Google’s latest model, shows a correct response rate to 41% of the questions in the Humanity’s Last Exam benchmark, a record in the industry.
- It is specifically intended for professionals such as scientists and mathematicians who require more sophisticated reasoning.
- Access to Gemini 3 Deep Think is limited to Google AI Ultra subscribers, with a subscription cost of 275 euros per month.
A cutting-edge language model
With the launch of Gemini 3 Deep Think, Google offers a language model that stands out for its ability to explore multiple hypotheses simultaneously. This advanced reasoning mode allows for even more sophisticated results, surpassing the performance of previous versions like Gemini 3 Pro.
Currently, Gemini 3 Deep Think is capable of correctly answering 41% of the questions in the Humanity’s Last Exam benchmark, a test known for its difficulty. This surpasses the 37.5% correct answers achieved by the previous model, Gemini 3 Pro.
Designed for experts
This model is not intended for the general public. It primarily targets researchers, developers, and mathematicians who need machines capable of thinking more deeply and with greater precision. Unlike other modes such as “Deep Research,” which rely on extended reasoning, Deep Think uses a completely distinct approach.
The IMO and ICPC technologies, recognized for their Olympic gold medals, are integrated into this model, thereby enhancing its capabilities to solve complex problems in mathematics and science.
Exclusive and costly access
Gemini 3 Deep Think is only accessible to Google AI Ultra subscribers, requiring a monthly subscription of 275 euros. This exclusivity is explained by the considerable resources the model consumes and its targeting towards a specific professional audience.
For regular users, Google continues to offer Gemini 3 Pro, while the Gemini 3 Flash version, optimized for faster queries, is in preparation.
Context and history of Google in AI
Since its inception, Google has established itself as a major player in artificial intelligence, investing heavily in the development of language models and other advanced technologies. The company has introduced several notable innovations, such as TensorFlow, an open-source library that has greatly facilitated the development of AI applications. With Gemini 3 Deep Think, Google continues to assert its leadership position in the field, meeting the growing needs of professionals for increasingly sophisticated solutions.