Google Slashes AI Prices, Rushes Out New Model

Google is rapidly advancing its AI capabilities with the release of Gemini 3.7 Flash, a new iteration of its "workhorse" model. This update emphasizes improvements in coding, agentic workflows, and knowledge work. The release follows a swift three-week development cycle, driven by developer feedback and algorithmic enhancements.

For developers, the significant news includes a temporary 50% reduction in API prices, effective until December 31, 2026. Input tokens will cost $0.75 per million, and output tokens $3.75 per million. These introductory rates aim to allow teams to evaluate the model's claimed cost-saving benefits in high-volume agent deployments.

The new model showcases enhanced abilities in adapting to roadblocks, clarifying intent, and executing multi-step plans with greater fidelity. Google reports substantial gains in software engineering benchmarks, particularly in production code quality and long-horizon evaluations, positioning 3.7 Flash as a competitive option in the AI landscape.

Gemini 3.7 Flash is accessible via the Gemini API, Google AI Studio, and Android Studio. Google is also implementing updated safeguards for chemical, biological, radiological, and nuclear risks, alongside cyber-offense misuse prevention.