On August 13, Google officially launched the new Gemini 3.7 Flash, just three weeks after the release of its predecessor, Gemini 3.6 Flash. Google positions it as the strongest main model to date for code writing and AI Agent scenarios, attributing it to significant breakthroughs at the algorithm level.
The model has seen significant improvements in software engineering and logic tasks: the score on the FrontierCode benchmark test increased from 34.4% to 43.6%, and the DeepSWE test score rose from 49.0% to 65.3%. It also performs well in areas such as web development, document understanding, and business process automation, with overall performance exceeding Claude Sonnet 5 and GPT-5.6 Terra.
In terms of ecosystem deployment and cost control, Gemini 3.7 Flash is now available through API, AI Studio, and Antigravity. Its initial pricing is set at $0.75 per million input tokens and $3.75 per million output tokens, which is a direct 50% reduction compared to the initial price of 3.6 Flash (the prices of both models are now equal).
This promotional pricing is expected to last until the end of 2026. With performance breakthroughs and reduced costs in complex coding and autonomous agent areas, high-performance large models are accelerating the commercialization of large-scale application deployments.
