Google is announcing a brand new Gemini mannequin at the moment, nevertheless it’s not the long-awaited 3.5 Professional. Gemini 3.7 Flash is now rolling out to interchange 3.6 Flash, which itself was launched only three weeks ago. This new “workhorse” mannequin is supposedly the product of core optimizations and developer suggestions, providing improved coding and agentic efficiency. And Google is hoping to counter the decrease price of some competing fashions with a decrease “introductory value” for 3.7 Flash.
In keeping with Senior Director Tulsee Doshi, Gemini 3.7 Flash is noticeably higher at coding than the earlier Flash launch. She cites a leap within the FrontierCode 1.1 Foremost take a look at from 34.4 to 43.6 p.c and DeepSWE v1.1 going from 49 to 65.3 p.c. As for the vibes, Gemini 3.7 Flash’s WebDev Area rating has risen to 1,588 from 1,538.
Folks turning to Gemini and hoping it can “know” issues might also see modest enhancements in Gemini 3.7 Flash. The GDP.pdf benchmark, which measures how properly a mannequin can course of complicated paperwork, has gone as much as 34 p.c versus 22 p.c with 3.6 Flash. AutomationBench exams how properly fashions can execute widespread enterprise workflows, and Gemini 3.7 Flash rose to 30.4 p.c from 3.6’s 17 p.c rating.
Credit score:
Google
Gemini 3.7 Flash is somewhat higher than 3.6 Flash in benchmarks.
Credit score:
These numbers definitely are greater. However are they sufficiently completely different to help a brand new mannequin launch simply three weeks after the final one? This can be extra about sustaining the looks of fixed enhancements in Google’s AI. All through 2024 and 2025, Google quickly made up floor to rival the very best AI popping out of competing AI labs. Issues seem to have slowed in 2026, although. At I/O in Might, Google promised that the flagship Gemini 3.5 Professional would launch in June, however that by no means occurred.
