Gemini 3.7 Flash: Google's Rapid AI Upgrade & Pricing

Google is rolling out a new Gemini model today, called 3.7 Flash. This new version replaces 3.6 Flash, which was released just three weeks ago. What's interesting is that this "workhorse" model aims to offer improved coding and agentic performance. Google is also introducing it with a lower "introductory price" to compete with other models. Senior Director Tulsee Doshi states that Gemini 3.7 Flash shows better coding abilities. For example, its score on the FrontierCode 1.1 Main test increased from 34.4 to 43.6 percent. Its DeepSWE v1.1 score also rose from 49 to 65.3 percent. For web development, the WebDev Arena score is now 1,588, up from 1,538. The model also shows modest improvements in processing complex documents, with the GDP.pdf benchmark rising to 34 percent from 22 percent. AutomationBench scores improved to 30.4 percent from 17 percent. The bottom line is that these updates suggest continuous, rapid advancements in AI technology.