Google unveils Gemini 3.7 Flash AI model for coding and agent workflows, but top AI model still delayed

Google unveiled its latest workhorse AI model, Gemini 3.7 Flash, on Thursday (Aug 13), but held off on announcing a release date for its most powerful AI model, Gemini 3.5 Pro. The new Flash model outperforms its predecessor in coding tasks such as debugging and generates deployable, production-ready code more efficiently on the first try, according to Google's blog post. Gemini 3.7 Flash can also complete app production in fewer prompts, improve the developer experience, and reduce the cost of tokens needed for operation.

In line with its commitment to AI safety, Google integrated stronger safety features into Gemini 3.7 Flash, including the ability to block malicious hacking and the misuse of dangerous chemical, biological, radiological, or nuclear materials without hindering safe and helpful work. Despite expanding its Gemini Flash lineup, Google has fallen behind rivals OpenAI and Anthropic in the race to develop the most advanced AI models.

This delay in Gemini 3.5 Pro has raised concerns among investors and questions about Google's ability to convert its massive AI investments into market-leading tools and services. CEO Sundar Pichai had previously stated that Google aimed to release models at a faster pace, and the company is already allocating significant computational resources to train its upcoming Gemini 4 model.