News in Short
- Google says Gemini 3.7 Flash is its most intelligent Flash model yet for coding and agents.
- The model improves coding, web development and multi-step workflow performance over Gemini 3.6 Flash.
- Gemini 3.7 Flash gets an introductory price of $0.75 per million input tokens and $3.75 per million output tokens.
- Gemini Spark will use Gemini 3.7 Flash to improve tool use and complex Google Workspace workflows.
Google has launched Gemini 3.7 Flash, its latest Flash-series model focused on coding, AI agents and complex knowledge-work tasks. The company says the model delivers major gains over Gemini 3.6 Flash while costing less during its introductory period.
Targets Coding and AI Agents
Google says Gemini 3.7 Flash delivers stronger performance across software engineering, web development and knowledge work.
For coding, the company reports higher first-pass accuracy and better performance in debugging and issue resolution. On FrontierCode 1.1 Main, Gemini 3.7 Flash scored 43.6 percent compared with 34.4 percent for Gemini 3.6 Flash. It also scored 65.3 percent versus 49 percent on DeepSWE v1.1.
Meanwhile, Google says the model can create more functional layouts and feature-complete web apps with fewer prompts. It scored 1588 on Arena.ai’s WebDev Arena, compared with 1538 for Gemini 3.6 Flash.
Lower Price for Developers
Google is also positioning the new model as a more cost-effective model for production AI agents. The model will be available at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens through the end of 2026.
Google says the model can better adapt when workflows hit roadblocks. It can also clarify intent, follow instructions more closely and spend more effort on multi-step planning and tool calls. That could reduce manual intervention and the number of retries required during complex development workflows.
Gemini Spark Gets the New Model
The new model will also power Gemini Spark from August 13. Google says the model will improve Spark’s knowledge-work capabilities and its use of Google Workspace tools. Spark can use the updated model to consolidate files, draft emails and update status documents. The model also includes updated safeguards covering cyber misuse and CBRN-related risks.
