AI Generated Image
What Google has teased about Gemini 4
9to5Google • • business
Key Points:
- Google is currently developing Gemini 4, described as its "most ambitious pre-training run yet," following the delay of Gemini 3.5 Pro's June launch.
- CEO Sundar Pichai emphasized Google's commitment to maintaining a leading position in AI development, particularly focusing on improving coding capabilities and agentic coding.
- Gemini 4 will require a larger base model to compete at the frontier of AI, with training already underway and an anticipated launch around November or December.
- Google prioritizes allocating Tensor Processing Units (TPUs) to support advanced AI development, ensuring resources are focused on competing at the frontier level.
- The company plans to continue frequent updates to Gemini 3.x Flash models, aiming for near-monthly releases with enhancements in agentic coding and other areas.
