Google released Gemini 3.7 Flash on August 14, saying the AI model improves coding tasks such as debugging, generates more deployable, production-ready code on the first try, completes app production in fewer prompts, and lowers token costs.

Google also said the model blocks malicious hacking and dangerous chemical, biological, radiological, or nuclear misuse.

The company did not provide a release date for Gemini 3.5 Pro after earlier delays.

Google has positioned Spark as an AI agent for consumers and small businesses that works across apps such as Gmail, Docs, and Maps.

Google said Gemini 3.7 Flash will power Spark from August 14.

The product is currently available to Google AI Ultra subscribers in select countries and some business users.

The update also comes as rivals push coding agents into developer tools.

Apple said Xcode will support OpenAI’s Codex and Anthropic’s Claude Agent for multi-step coding tasks.

Google’s wider frontier AI safety framework also covers cyber and chemical, biological, radiological, and nuclear risk thresholds, alongside deployment mitigations.

That provides context for the new blocking claims.

🔗 Source: Bloomberg

Recent Google developments