Ready to start building with EmbeddingGemma 2? Our Developer Guide walks you through embedding multimodal content across text, code, images, video, and audio using the sentence-transformers library and more! The workflow:

Finding the exact DNA that makes crops climate-resilient used to take years of physical breeding.

Introducing EmbeddingGemma 2! 🚀 Our lightweight, multimodal embedding model maps text, code, images, video, and audio into a single, unified embedding space. Optimized for on-device use cases, it features:

What if every student had the superpower of a personal tutor?

Turn DiffusionGemma into a Jev-like model.

Today’s best coding agents rely on massive API models running in the cloud. What if every developer could rely on an autonomous agent equally capable offline, on consumer hardware? We’re challenging you to close the gap. The Gemma 4 Developer Agent Competition with @Kaggle is live!

Announcing the Google Gemma Hardware Hackathon at the legendary AGI House! Join us for the builder finale to Open Source AI Week. We’re hosting an intensive day dedicated to the future of physical AI. Apply now to get hands-on with the latest mobile, embedded and robotics hardware x Gemma models.

Build agentic workflows completely offline. The Antigravity SDK now supports local execution with Gemma 4 and LiteRT. Run agents entirely on your local machine with: 💵 Zero token costs 🔒 Total data privacy 🔌 Offline reliability

Deploying DiffusionGemma-Jev (djev) just got a lot easier. You can now spin up a Jev API-compatible endpoint on Google Cloud Run using a single command. Performance is solid: ~35-60 ms for single step latency and batch@32 is ~100-123 requests/sec.

"DiffusionGemma as Jev" showcases the power of non-autoregressive architectures. While Jev demonstrates the value of rapid decision models, running DiffusionGemma in this paradigm leverages canvas diffusion to evaluate structured choices in a single parallel pass:

We prototyped a tactile desk robot using LEGO, a Raspberry Pi, and a hybrid LLM setup: ✨ Gemma 4 runs locally for instant, private zero-cost chat ✨ Auto-escalates to Gemini Flash for complex reasoning & code

We prototyped, a tactile desk robot using LEGO, a Raspberry Pi, and a hybrid LLM setup: ✨ Gemma 4 runs locally for instant, private zero-cost chat ✨ Auto-escalates to Gemini Flash for complex reasoning & code

Looking for a better local model that runs on a MacBook Air? Our latest update brings full Gemma 4 12B support to Google AI Edge Gallery app on Mac. ✨ Multimodal: Vision & Audio inputs 🎛️ Configurable vision token budgets ⚡️ MTP for blazing-fast inference 💻 Optimized to run on 16GB Macs!

Standup Pulse: an open-source project running async Slack standups using Gemma 4 26B-A4B (GGUF via llama.cpp) on an Apple M5 Max.

Running AI in space requires strict power and bandwidth management.

Sign language AI should not depend on Wi-Fi or expect Deaf folks to do all the adapting. Meet KawanIsyarat, a local-first mobile tool bridging BISINDO and spoken Indonesian offline using Gemma 4.

You don't need to be a command-line expert to run local AI.

Stitching AI video is easy; keeping the story coherent is hard.

Deploying AI at the edge makes local public service tools practical and accessible.

Gemma is only as powerful as the community building with it. We launched awesome-gemma to highlight amazing community-made tools, research, and applications in the ecosystem. What have you built? Drop a link below or submit a PR to get featured: https://github.com/google-gemma/awesome-gemma

Great news for app developers who want to code on the go! Gemma 4 is now the recommended local model provider in Android Studio! Make code changes completely offline in agent mode with zero token quotas. With a local AI coding agent, your code never leaves your machine!

Running Gemma 4 26B A4B on a Mac just got 2x faster! The developer community has been grinding on the https://mlx.fast leaderboard, pushing Apple Silicon to its limits.

The latest technology should be accessible to the elderly, empowering them to navigate a changing world on the devices they already own. (1/3)

Gemma 4 E2B running fully locally on an 8GB NVIDIA Jetson Orin Nano. Using native image/audio input and function calling, this edge setup is capable of vision, voice interaction, and camera control, all within an 8GB memory footprint!

Addressing the world's most pressing challenges requires robust solutions. These Gemma 4 Good projects fine-tuned models to address key challenges in anticipating and mitigating problems that affect our environment and society. (1/4)

AcuIfero4Vigia

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

youtube.com

3,431 tokens/s on Gemma 4 31B! Recent benchmarks on NVIDIA's Groq 3 LPX accelerator show a median output of ~3,400 tokens/s across both 10k and 100k input sequences, the fastest 100k-context speed recorded for the model to date.

Fast and local voice typing with Gemma 4. Phonon demonstrates the utility of lightweight, on-device models for privacy-preserving workflows. This open-source voice typing tool runs Gemma 4 (E2B 4-bit) locally via MLX to process audio and screen context without cloud routing.

The wait is over. 🏆 We challenged you to use Gemma 4 to tackle real-world problems for the Gemma 4 Good Hackathon. Your submissions set a new standard for technical excellence and real-world impact. Meet the innovators who took home the top prizes.

You don't always need a frontier model. A recent benchmark found that Gemma 4 31B matches Sonnet 5 on answer quality at ~40x lower cost. With high cost-efficiency and low latency, Gemma unlocks high-volume use cases that are uneconomical with larger models.

Tonight we’re gathering in SF to celebrate 1 BILLION Gemma downloads! While we can only fit a few of you in the room, we’ll be raising a glass to the millions of developers worldwide who are driving the Gemmaverse forward. 🥂 (1/2)