🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Gemini 2.5 Flash-Lite independent benchmarks: ~5x cheaper and ~1.7x faster than regular Flash but with reduced intelligence Google Gemini 2.5 Flash-Lite Preview (06-17) is an upgraded version of Gemini 2.0 Flash-Lite released in Feb 2025. Gemini 2.5 Flash-Lite is a hybrid model,

We used native audio in the Gemini LIve API + Pipecat AI + Supabase to upgrade kwindla 's personal todo app with generative UI. The assistant created 🥦 a grocery list 🥸 a reading list including looking up the names of author and books in a trilogy 👨‍🌾 a work todo list with

A Benchmark for Deep Research Agents? 👀 DeepResearch Bench is a new benchmark of 100 PhD-level research tasks across 22 distinct fields to systematically assess the report generation quality and citation accuracy of deep research agents. Benchmark Creation: 1. Collected over

The Gemini Live API now supports 2.5 Flash (gemini-live-2.5-flash-preview) as well as part of the cascaded architecture. Developers will likely see improvements in response accuracy, creativity, instruction following and especially function calling. Pricing for this model on

This is important. Please note if you are still using Pro 0506 or Pro 0325

Gemini 2.5 is driving new advancements in robotics with its strong coding, spatial reasoning, and multimodal capabilities. 🦾

Imagen 4 is now available in the Gemini API and Google AI Studio! 🐎 Imagen 4 is the goto model for most tasks 🐕‍🦺 Imagen 4 Ultra is for more precise instruction following Congrats Alisa Fortin Seth Odoom, Guillaume Vernade and team developers.googleblog.com/en/imagen-4-no…

This is emotional for me having built the first version of the data science agent that got released in Colab along with Chris Perry Kath Korevec Vincent Nallatamby Jane FIne and others. And that was about a year ago. Incredible how far the team has come since then! Congrats!

Final reminder! This time we made sure to give lots of notice 😇 We will be redirecting "gemini-2.5-pro-05-06" and "gemini-2.5-pro-03-25" endpoints to our GA model gemini-2.5-pro on tomorrow Jun 26 between 10-11 am PST

Google's AI domains are thriving. Month-over-month traffic growth in May: Google DeepMind: 4,258,300 → 11,189,832 (+162.78%) Google Labs: 9,796,459 → 19,792,943 (+102.04%) Google AI: 1,349,027 → 2,042,165 (+51.38%) NotebookLM: 48,340,200 → 64,906,295 (+34.27%) Google Gemini App:

I'm really impressed by the new Gemma 3n I tried a 7.5GB model from Ollama and a 15GB model through mlx-vlm - they seem very capable, and this is the first model of that size I've tried that can handle both image AND audio input in addition to text! simonwillison.net/2025/Jun/26/ge…

Hard to believe that the 1st Gemma model was only launched early 2024. And now Gemma 3n (powered by the Matformer architecture) achieves 1300+ on LMArena! Available in 2sizes based on effective parameters: E2B (actual 5B) and E4B (actual 8B). Architectural innovations allow them

Favorite use case of the Gemini CLI so far! Use it to query repos. Here I use it on github.com/google-gemini/… and ask questions ❓How do I get an API key for the Gemini API and where do I set it up in this project? ❓Walk me through the main Flask application code - what does

wow

Gemini CLI Update from last week! We merged 85 PRs from 51 unique contributors. Here are key improvements: - Gemini CLI can now use audio and video (santhoshkumarCodes) - Upgrade to Ink 6 and React 19 (SandyTao520) - GEMINI md can import other markdown files with @.

Happy 4th everyone! Gemini app helped me come up with this recipe "I have red potatoes, leeks, fiddle head ferns, shallots and parsley. Help me make a recipe that's a great side dish for crab cakes for a 4th of July dinner. The constraints are that I cannot use dairy or

We now support batch mode in the Gemini API for use cases such as bulk content gen and processing and evals! 🎏Higher throughput and rate limits - you can submit large jobs, offload the scheduling and processing, and retrieve your results within 24 hours 💰 50% discount compared

🖊️You can now improve function calling performance in the Gemini API by using thought signatures which ensure that the model maintains context across multiple turns. 🧩Thought signatures are encrypted representations of the model's internal thought process. The model returns