Google Gemma

Google AI Pro plans now package premium Colab benefits, landing you A100s with 80GB VRAM (Ultra subs even unlock H100s 🔥).

Here’s what that 80GB headroom unlocks directly in your browser:

✨ Run Gemma 4 31B in bf16
✨ Full fine-tune up to Gemma 4 E4B in bf16
✨ LoRA tune up to Gemma 4 26B A4B in bf16
✨ QLoRA tune up to Gemma 4 31B in bf16

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:

1️⃣ Load the model (via your preferred framework)
2️⃣ Embed text & code with task prompts
3️⃣ Embed images, video, audio & interleaved inputs
4️⃣ (Optional) Truncate dimensions with Matryoshka to save space

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

To speed this up, AI lab Living Models pairs Gemma 4 E4B with BOTANIC-1, a genomic language model. Gemma prepares the genomic data, feeding filtered candidates to BOTANIC-1 to score evolutionary impact and pinpoint the causal mutation.

In a recent test, they pinpointed a target melon yield mutation in an afternoon of computation, ranking it #1 out of 2,494 possibilities.

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:

- 740M parameter form factor with modular encoders
- Flexible dimension sizes (768dim-128dim) via Matryoshka Representation Learning (MRL)
- 8K context window (4x larger than text-only EmbeddingGemma)
- A commercially permissive Apache 2.0 license

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

New research shows AI delivers immediate GRE learning gains equivalent to expert humans, with Gemma 4 31B leading the way. It serves as a valuable resource to assist educators, helping experts scale their impact and making guidance accessible to everyone.

Turn DiffusionGemma into a Jev-like model.

This brings a new level of precision to structured tasks. By seeding a canvas with a response template, vLLM lets you extract confidence scores and probability distributions for yes/no, multiple-choice, and scored questions, all in a single denoising step.

Fast, efficient, and ready to scale!

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!

💰 Total Prize Pool: Over $110k
⏰ Entry Deadline: Nov 2, 2026

Accept the challenge and get the starter kit:
kaggle.com/competitions/g…

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.

The only rule? No cloud anywhere in the loop. Let’s build the future of embodied intelligence together.

📅 Oct 24 | 📍 Hillsborough, CA

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

Bonus feature: Support for OpenAI-compatible endpoints. Use Ollama, llama.cpp, vLLM and more to serve Gemma 💪

Get started: pip install google-antigravity litert-lm

Read the details: developers.googleblog.com/introducing-su…

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.

It's a straightforward way to experiment without needing your own GPU. Runs at roughly $3/hr and drops to $0 when idle.

Get the code and instructions here: github.com/taeold/djev-run

"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:

⚡ ️Massive Parallelism: Denoises across an open canvas in a single step instead of sequential autoregressive token generation (~0.2s on a DGX spark).

🧠 Full Bidirectional Attention: Allows every option to attend to the full context concurrently, yielding well-calibrated decision distributions.

👁️ Multimodal Grounding: Inherits Gemma 4's spatial vision capabilities for complex visual and text decisions.

Read more about this approach here:
github.com/vllm-project/v…
x.com/mmastrac/statu…
x.com/mmastrac/statu…

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

Here’s how we wired and coded DinoDesk AI:
dev.to/googleai/how-w…

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!

Download it today: ai.google.dev/edge/gallery

Standup Pulse: an open-source project running async Slack standups using Gemma 4 26B-A4B (GGUF via llama.cpp) on an Apple M5 Max. It pairs Mastra for typed tool selection with CopilotKit Channels for restrained Slack Block Kit (Slack's native layout format) interactions while keeping all model inference, standup records, and traces local in SQLite.

This architecture is a great demonstration of how to connect Slack to a project without exposing the local server!

🔗 Blog: soverius.ai/blog/building-…
🔗 Repo: github.com/Soverius-AI/st…

Running AI in space requires strict power and bandwidth management.

Satlyt is deploying Gemma directly on satellites to analyze telemetry and diagnose faults locally. Running Gemma 3 1B onboard cuts diagnostic data payloads by over 64%, turning raw system logs into compact summaries before they are sent back to Earth.

Next launch target: Gemma 4 E2B on NVIDIA Jetson Orin Nano to provide further reasoning and vision capabilities at the edge… of space.

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.

Gemma 4 handles all vision, audio, and text processing directly on-device. Congrats to the team on their Cactus category win for the Gemma 4 Good Hackathon.

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

llama.app wraps the heavy lifting of llama.cpp into a clean UI, making it simple to deploy and manage models like Gemma 4. Instead of dealing with config files, you get one-click model downloads, clear memory estimates, and no coding required!

Watch the video to see Gemma 4 in action as it parses tables from receipts, streams reasoning logs, and connects to MCP for web search.

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

This new ComfyUI sampler enables longer videos by processing them in chunks (saving VRAM). To keep scenes consistent, Gemma 4 runs under the hood to write the continuous prompts to keep characters, backgrounds and camera movements consistent between chunks.

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

PreVillage connects citizens in Nepal to government services over WhatsApp. Powered by Gemma via llama.cpp on a Raspberry Pi, it cleans raw speech transcripts, reasons through user intent, and retrieves verified public sources directly on-device.

Congrats on the llama.cpp prize winner for the Gemma 4 Good Hackathon!

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: github.com/google-gemma/a…