Gemini 3.7 Flash Intelligence Index Rises 4 Points in Three Weeks

Google releases Gemini 3.7 Flash with 4-point IQ gain and 3x speed boost over rivals. Programming scores jump significantly, while input costs are halved temporarily until year-end.

Woofun AI reports that Google has launched Gemini 3.7 Flash, marking a third-generation update within a three-week window following the 3.6 Flash release. The model is now accessible via the Gemini API, AI Studio, Antigravity, and Gemini Spark, maintaining a primary focus on programming and Agent capabilities.

Artificial Analysis data indicates the new model achieved an intelligence index of 56 in the high-inference tier, a 4-point increase over its predecessor. This places it 1 point behind GPT-5.6 Terra and Muse Spark 1.2, while delivering generation speeds of approximately 340 tokens per second—nearly triple that of GPT-5.6 Terra. Internal testing shows DeepSWE v1.1 scores rising from 49.0% to 65.3%, and AutomationBench enterprise automation scores increasing from 17.0% to 30.4%. Google attributes these gains to algorithmic enhancements. Pricing is temporarily reduced to $0.75 per million input tokens and $3.75 per million output tokens through the end of the year, reverting to $1.50 and $7.50 respectively on January 1, 2027. No official release date has been set for Gemini 3.5 Pro.