This month’s Gemini updates feel like Google gave the app a proper tune-up, then added a few clever shortcuts on top. The big theme is stronger reasoning, better multimodal support, and more useful ways to get from a prompt to something actually handy. If you use Gemini for writing, research, planning, or dev work, there’s a decent chance one of these changes will save you a few minutes and a small amount of brain ache.
Below I’ve pulled together the most recent updates from the last 14 days, keeping it simple and practical.
Gemini 3 Flash is now the default model
Gemini 3 Flash is now the default model in the Gemini app, and Google says it’s a major upgrade over Gemini 2.5 Flash. It’s designed to be fast, but it also handles more complex questions and mixed inputs like images, audio, and text more capably than before.[1]
Why it matters: you should notice better everyday responses without having to fiddle around with model choices. That’s handy if you’re drafting campaign briefs, turning a photo into notes, or asking Gemini to make sense of a messy spreadsheet screenshot.
For developers, this means a stronger baseline model for multimodal workflows. For marketers and writers, it means quicker first drafts and better support when you’re working with visual references, voice notes, or a pile of half-baked ideas from a Monday morning meeting.
Gemini 3 Pro is rolling out to everyone
Gemini 3 Pro is now rolling out to all Gemini app users, with higher limits for Google AI Plus, Pro, and Ultra subscribers.[1]
This matters because it gives more people access to Gemini’s stronger reasoning model without needing to hunt through a bunch of settings. If you’re doing deeper analysis, comparing product positioning, or asking for a structured breakdown of customer feedback, Pro should be the one to try.
In real-world terms, this is the model I’d reach for when I need a cleaner answer to “what’s the pattern here?” rather than just “what’s this?” It’s useful for analysts reviewing call transcripts, researchers organising notes, and founders pressure-testing ideas before they ship.
Gemini 3 Deep Think is now available for Ultra subscribers
Google AI Ultra subscribers can now use Gemini 3 Deep Think in the Gemini app.[1]
Deep Think uses iterative reasoning and multiple hypotheses to work through harder problems, especially in maths, science, and logic-heavy tasks.[1]
That makes it useful when you’re stuck on a knotty problem and need a model that slows down a bit and thinks it through properly. Picture debugging a stubborn workflow, checking assumptions in a technical plan, or working through a forecasting question that has more moving parts than your usual Tuesday to-do list.
For teams, this is the “let’s not rush this” mode. It’s not for every quick prompt, but when the stakes are high and the answer needs more depth, this is the update worth knowing about.
Gemini 2.0 Flash Thinking has been improved
An improved version of Gemini 2.0 Flash Thinking is now available to Gemini app users.[1]
Google hasn’t positioned this as a flashy new toy so much as a smarter version of an already useful reasoning model. That usually means better performance on prompts that need a bit more care, without making you wait around forever.
This is especially useful for people who want faster reasoning support for tasks like summarising a client call, untangling a project brief, or checking whether a piece of copy actually matches the ask. It’s the kind of update you may not notice immediately, which is often a good sign.
Gemini can now use your Search history for personalised answers
Gemini can now use your Search history to give more personalised responses, if you choose the experimental Personalization model from the dropdown.[1]
Google says Gemini will analyse your prompt and connect it with your past searches when that helps improve the answer.[1]
That can be handy if you’ve been researching a topic over several sessions and want Gemini to stay in the loop instead of starting from scratch every time. For example, if you’ve been comparing Shopify apps, planning a campaign, or looking into a niche industry trend, Gemini can potentially give you answers that feel more connected to your actual work.
There’s a practical catch here, of course: it’s an opt-in experimental feature, so it’s best used when personal context is genuinely helpful rather than when you want a fully fresh take.
Richer, more interactive responses are now part of Gemini
Gemini can now create richer, more dynamic responses for certain topics, including interactive explorations, layered images, timelines, and narrated video overviews.[1]
Google says Gemini will choose the best way to organise information for the prompt, which is a nice way of saying it won’t just dump a wall of text on you when a visual answer makes more sense.[1]
This is great for people who learn better by seeing how things connect. A researcher could use it to unpack a timeline. A marketer could use it to explore product messaging visually. A developer could use it to understand a complex system diagram without squinting at a static screenshot like it’s 2014.
It’s still early days for this style of response, but it points to Gemini becoming more useful for explanation, not just generation.
If you want to see what’s new for yourself, head over to Gemini, try the latest models, and kick the tyres a bit. If you rely on Gemini for work, keep an eye on the release notes and send feedback when something’s brilliant or a bit clunky. That’s usually how the good stuff gets better.
