Google announced Gemini 3.7 Flash on August 13, 2026, positioning it as an advancement in the Flash model series optimized for coding and agentic workflows. The release follows Gemini 3.6 Flash by three weeks and incorporates developer feedback alongside algorithmic innovations.

Performance Improvements

3.7 Flash shows measurable gains in code generation and debugging. On FrontierCode 1.1 Main, it achieved 43.6% accuracy versus 34.4% for 3.6 Flash. On DeepSWE v1.1, performance reached 65.3% compared to 49.0% previously. The model demonstrates higher first-pass code accuracy and improved capability to produce production-ready code with fewer manual interventions.

For web development tasks, 3.7 Flash generates more functional layouts and complete applications in fewer prompts. Arena.ai's WebDev Arena benchmark shows an Elo rating of 1588 for 3.7 Flash versus 1538 for 3.6 Flash. The model excels at UI generation with strong design adherence whether working from screenshots, images, or design systems.

In knowledge-heavy domains like finance, law, and biosciences, 3.7 Flash improved reasoning and accuracy. On the GDP.pdf benchmark for complex document processing, it achieved 34.0% versus 22.0% for 3.6 Flash. AutomationBench results showed 30.4% versus 17.0%, demonstrating effectiveness at real-world business workflow automation.

Developer Experience

3.7 Flash exhibits better instruction-following and clearer intent comprehension than its predecessor. The model approaches multi-step planning more deliberately, reduces manual oversight requirements, and requires fewer retries in engineering workflows. Early customer feedback highlights performance gains and precision improvements.

Pricing and Availability

Introductory pricing through December 31, 2026 is set at $0.75/1M input tokens and $3.75/1M output tokens—half the original 3.6 Flash cost per million tokens. From January 1, 2027, pricing will increase to $1.50/1M input and $7.50/1M output tokens.

Developers can access 3.7 Flash via Google Antigravity, Gemini API through Google AI Studio, and Android Studio. Enterprise customers have access through the Gemini Enterprise Agent Platform and Enterprise app. Individuals can use the model via Gemini Spark, a 24/7 personal agent available to Google AI Pro and Ultra subscribers in over 160 countries.

Gemini Spark Integration

The personal agent Gemini Spark now uses 3.7 Flash to improve efficiency for knowledge work with enhanced tool integration for Google Workspace applications. This enables better handling of complex multi-skill workflows including file consolidation, email drafting, and status document updates.

Safety

Gemini 3.7 Flash includes updated Frontier Safety safeguards covering Chemical, Biological, Radiological, and Nuclear (CBRN) domains and cyber offense capabilities, aligned with Google's bioresilience and cyber programs.