Google could well redefine the landscape of artificial intelligence with the arrival of Gemini 3.0. While its predecessor, Gemini 2.5 Pro, continues to dominate the rankings, the tech community is buzzing with excitement at the possibility of an imminent announcement. What are the mysteries surrounding this new model, and what does it mean for the future of AI?
3 things you need to know
- The “cheetah” model may be linked to Google, although it is being released by an undisclosed provider.
- Unlike traditional models, “cheetah” uses diffusion technology to generate text.
- Gemini 3.0 could be announced as early as October 2025, potentially surpassing GPT-5 and Claude Sonnet 4.5.
The mystery of the “cheetah” model
A model called “cheetah,” available since October 5 on the Cursor platform, is sparking much speculation. Although its creator has not yet been officially identified, many believe it could be Google. This model stands out for its exceptional speed, charging $1.25 per token mission input and $10 per output, which is unusual for a test model.
Understanding diffusion technology
Unlike traditional regression models that generate text sequentially, diffusion technology starts with a noisy sequence that it gradually refines. This approach allows large blocks of text to be generated at impressive speeds, which is ideal for applications requiring rapid iterations, such as coding and mathematics.
Gemini 3.0: a potential breakthrough for Google
With a possible announcement in October 2025, Gemini 3.0 could mark a major breakthrough for Google in the field of AI. The model could incorporate innovations such as those seen in “cheetah,” offering performance that could rival that of OpenAI’s GPT-5 and Anthropic’s Claude Sonnet 4.5.
Google and the evolution of its AI models
Since its inception in the technology sector, Google has constantly pushed the boundaries of innovation. With the launch of Gemini 2.5 Pro, the company has consolidated its position as a leader in artificial intelligence. Speculation about Gemini 3.0 shows the expectation of another significant breakthrough, reinforcing Google’s commitment to transforming AI into an even more powerful and accessible tool.
At the same time, Google DeepMind, a Google subsidiary specializing in artificial intelligence, has been at the forefront of research on diffusion models. The development of Gemini Diffusion, unveiled at Google I/O, demonstrated the company’s commitment to integrating cutting-edge technologies into its products. If “cheetah” is indeed a Google product, it would represent a significant step forward in the evolution of diffusion technology applied to AI.