Say hello to Grounding with Google Search, available in the Gemini API + Google AI Studio! You can now access real time, fresh, up to date information from Google Search when building with Gemini by enabling the Grounding tool. developers.googleblog.com/en/gemini-api-…

We’ve been *thinking* about how to improve model reasoning and explainability Introducing Gemini 2.0 Flash Thinking, an experimental model trained to think out loud, leading to stronger reasoning performance. Excited to get this first model into the hands of developers to try

Lateral thinking: "Wait a minute... could I turn one of the numbers upside down?"

We released Gemini 2.0 Flash Thinking today! ⚡️🤔 It's a small step towards improved reasoning via inference-time compute, built on top of our small and mighty 2.0 Flash!

I'm pretty amazed by how easy it is to localize frontend apps with AI agents. I had a meeting with people in Japan, and wanted to create a prototype of our app in Japanese. In about 8 hours, OpenHands generated 6200 lines of code and now our app is localized in 10 languages.

A new year, a new challenge. I recently joined AI at Meta to improve evaluation and benchmarking of LLMs. I'm excited to push on making LLMs more useful and accessible, via open-sourcing data/models and real-world applications. I'll continue to be based in Berlin.

gemini 2.0 pro 😃

1/ Today we are releasing Gemini 2.5 Pro Experimental, our newest Gemini model with integrated “thinking” and significant performance gains. Very proud of the whole team! 🧵

Introducing Gemini 2.5 Pro Experimental! 🎉 Our newest Gemini model has stellar performance across math and science benchmarks. It’s an incredible model for coding and complex reasoning, and it’s #1 on the lmarena.ai leaderboard by a drastic 40 ELO margin. Only a handful of

Gemini 2.5 Flash is here, our first unified reasoning model with thinking budgets. 🔥 It’s on the perato frontier and punches above its price and size!! developers.googleblog.com/en/start-build…

⚡Gemini 2.5 Flash is now in Preview. Available on Google AI Studio, it’s our first fully hybrid reasoning model that lets you toggle thinking or set budgets for the optimal quality/cost/latency mix. Maintain 2.0 Flash speed + improved perf even when thinking is off. →

Excited to introduce Gemini 2.5 Flash our most cost-efficient thinking model. We are once again at the frontier here. Pretty good well rounded performance.

We've just given our most powerful workhorse model a big upgrade to Gemini 2.5 Flash. You can try it now in preview on ai.dev - yet another Gemini data point on the cost-performance pareto frontier!

Competing with ourselves is getting a bit boring

Gemini 2.5 Pro shipping a granular thinking budget that works could actually be a pretty big deal. The glorious slider that comes before the model just knows how much thinking to do. Helps limit overthinking, helps collect good data on "how hard users think the model is"

Our latest Gemini 2.5 Pro update is now in preview. It’s better at coding, reasoning, science + math, shows improved performance across key benchmarks (AIDER Polyglot, GPQA, HLE to name a few), and leads lmarena.ai with a 24pt Elo score jump since the previous version. We also

Our latest update to Gemini 2.5 Pro is here. It's SoTA on GPQA Diamond, AIDER and HLE. The team has also worked hard to improve the model on style, persona and creativity. We're excited to see what you build with it. Please let us know any feedback as we're eternally cooking.

Fiction.live We like long context. Go beyond 192k plz ; )

Gemini 2.5 Pro 06-05 has set a new SOTA on the aider polyglot coding benchmark, scoring 83% with 32k thinking tokens. The default thinking mode, where Gemini self-determines the thinking budget, scored 79%. Full leaderboard: aider.chat/docs/leaderboa…

Made it to no.1 in the App Store. Congrats to the Google Gemini App team for all their hard work, and this is just the start, so much more to come!