Google has released Gemini 3.7 Flash today, replacing its predecessor Gemini 3.6 Flash just three weeks after that release.
The new model shows measurable improvements in coding (FrontierCode 1.1 Main rose to 43.6 percent from 34.4 percent, and DeepSWE v1.1 rose to 65.3 percent from 49 percent) and business-workflow tasks (AutomationBench rose to 30.4 percent from 17 percent), and Google is offering it at a lower introductory price to compete with cheaper rival models.
The rapid succession suggests Google is working to demonstrate continuous progress after a slower period in 2026, though it still has not delivered the long-promised Gemini 3.5 Pro.
What happened
Google announced Gemini 3.7 Flash today, rolling it out to replace Gemini 3.6 Flash, which was released only three weeks ago. The new model shows improved coding and business-task performance across multiple benchmarks: FrontierCode 1.1 Main improved from 34.4 to 43.6 percent, DeepSWE v1.1 from 49 to 65.3 percent, WebDev Arena from 1,538 to 1,588, GDP.pdf from 22 to 34 percent, and AutomationBench from 17 to 30.4 percent.
Why it matters
Google is offering Gemini 3.7 Flash at a lower introductory price to compete with lower-cost rival models. The rapid succession of releases—two Flash versions in three weeks—suggests Google is working to maintain momentum in AI development after a slower period in 2026; the company had promised Gemini 3.5 Pro would launch in June at its May I/O event, but that did not happen.
What to watch
Whether the benchmark gains justify the fast release cycle, and whether Gemini 3.5 Pro (the long-awaited flagship model) arrives soon. The introductory pricing for 3.7 Flash has not been specified in detail.
Google announced Gemini 3.7 Flash today as a replacement for Gemini 3.6 Flash, which had been released just three weeks prior. According to the company, the new "workhorse" model is the result of core optimizations and developer feedback, with a focus on improved coding and agentic performance—the ability to take actions on behalf of users in workflows.
Senior Director Tulsee Doshi highlighted the coding improvements in particular. On the FrontierCode 1.1 Main benchmark, Gemini 3.7 Flash scored 43.6 percent, up from 3.6 Flash's 34.4 percent. On DeepSWE v1.1, the new model reached 65.3 percent compared to the previous 49 percent. For general task performance, WebDev Arena scores rose to 1,588 from 1,538. When processing complex documents—measured by the GDP.pdf benchmark—3.7 Flash achieved 34 percent versus 22 percent for 3.6 Flash. On AutomationBench, which tests how well models execute common business workflows, Gemini 3.7 Flash scored 30.4 percent, a significant jump from 3.6's 17 percent.
Google is attempting to counter lower pricing from competing AI labs by offering Gemini 3.7 Flash at a lower introductory price, though the exact pricing details were not announced. The rapid release cycle—two Flash versions in three weeks—reflects a broader pattern in which Google has worked throughout 2024 and 2025 to close the gap with leading competitors. However, the article notes that progress appears to have slowed in 2026. Google had promised at its May I/O event that Gemini 3.5 Pro, the flagship model, would launch in June, but that timeline was not met. The quick succession of Flash releases may indicate Google is maintaining momentum in the interim by demonstrating incremental progress rather than waiting for the more substantial 3.5 Pro release.
Google's release of Gemini 3.7 Flash represents a notably accelerated development cycle. The model replaces 3.6 Flash after only three weeks, a timeframe that contrasts sharply with typical major-model release schedules in the AI industry. The performance gains, while measurable across multiple benchmarks, are modest in absolute terms—for instance, AutomationBench rose from 17 to 30.4 percent, and GDP.pdf from 22 to 34 percent. The body suggests this rapid succession may reflect Google's strategic goal of signaling continuous innovation in AI, particularly following a slower period in 2026. The company had publicly committed to launching Gemini 3.5 Pro in June at its May I/O event, but that milestone was not met, which may explain why the company is emphasizing incremental improvements to the Flash line. By pricing 3.7 Flash at an introductory discount, Google appears to be responding to competitive pricing pressure from rival AI labs, even as it continues work on the flagship 3.5 Pro model.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
IBM and OpenAI announced a strategic partnership on August 13, 2026, embedding OpenAI frontier models like GPT…
Apple is in talks with publishers to pay them for content licensing, as the company works to improve its AI-po…
CrowdStrike Holdings (NASDAQ:CRWD) posted Q2 2026 results with Net New Annual Recurring Revenue of $256M (up 3…
S&P Global expanded its partnership with Microsoft to integrate its financial, company, and energy intelligenc…
Pfizer has developed a federated AI pipeline that enables multiple parties to collaborate on machine learning…
Hitachi and SZTAKI have developed a technology that links generative AI with mathematical optimization to auto…
The AI news that matters, in one minute each morning.
Sign up free