Pinned Tweet
Hoy es un buen día para preguntarnos por enésima vez qué pasa con las mujeres y la informática politikon.es/2015/06/30/las-… #WomenInSTEM
Si estudias en España, echa un vistazo: Google AI Plus for Students
goo.gle/Gemini_edu_sr
Sabela retweeted
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
Sabela retweeted
The technical report for Gemma 4 is now live on Arxiv: arxiv.org/abs/2607.02770
💎💎💎💎
Sabela retweeted
Want to host your own Gemma hackathon?
We’re sponsoring 1-day hackathons on Kaggle to help developers dive into open models! From building lightweight tools to tackling your community's unique challenges, this is your chance to lead the charge with Gemma 4.👇
Meet DiffusionGemma ⚡ Our latest experimental open model (Apache 2.0) that generates text up to 4x faster.
Instead of predicting and typing just one word at a time like most language models, it drafts and refines entire blocks of text simultaneously.
Here’s how it works 🧵 ↓
Sabela retweeted
All parameters used for all modalities!
Really excited and proud to release this competitive multimodal 12B model that does not need extra encoders.
Sabela retweeted
Simplicity + Performance. Impressed by the work of the team to remove all the encoders from a multimodal Transformer and keep its quality.
Sabela retweeted
Our goal in the Gemma team is to ship models that are useful by generalizing to unseen tasks. Hence, we are extremely strict about not doing anything that would target specific benchmarks instead of teaching the model broad capabilities. It's great to see this reflected here.
Qwen3.5 27B vs Gemma 4 31B IT
Here is why I think Gemma 4 31B wins:
Benchmark scores of both models are very close.
Generalization is not.
I reused NVIDIA’s CoDeC metric to probe "benchmark affinity".
Officially, CoDeC > 80 suggests contamination.
I've never seen this. But even below that, it can still reveal how benchmark-affine a model is.
Higher CoDeC scores mean the model is more "comfortable" with the benchmark.
So the ideal profile is:
high benchmark scores + low CoDeC.
By that lens, Gemma 4 looks clearly stronger than Qwen3.5.
Gemma 4 likely generalizes better to unseen tasks.
I observed the same pattern earlier with Gemma 3 vs Qwen3.
I wrote about this a few months ago:
kaitchup.substack.com/p/did-…
Sabela retweeted
Really proud of everyone in the team that worked hard for this release: Gemma 4 is here! blog.google/innovation-and-a…
Sabela retweeted
Meet Gemma 4!
Purpose-built for advanced reasoning and agentic workflows on the hardware you own, and released under an Apache 2.0 license.
We listened to invaluable community feedback in developing these models. Here is what makes Gemma 4 our most capable open models yet: 👇
Sabela retweeted
Excited to launch Gemma 4: the best open models in the world for their respective sizes. Available in 4 sizes that can be fine-tuned for your specific task: 31B dense for great raw performance, 26B MoE for low latency, and effective 2B & 4B for edge device use - happy building!
Today we're releasing Gemma 4, our new family of open foundation models, built on the same research and technology as our Gemini 3 series. These models set a new standard for open intelligence, offering SOTA reasoning capabilities from edge-scale (2B and 4B w/ vision/audio) up to a 26B parameter MoE model and a 31B dense model. By releasing Gemma 4 under the Apache 2.0 license, we hope to enable more innovation across the research and developer communities. Our earlier Gemma 3 models were downloaded 400M times and over 100,000 variants of those models have been published, so we're excited to see what the community will do with the even better Gemma 4 models!
Learn more at blog.google/innovation-and-a… and goo.gle/gemma-4-apache-2
Great work by everyone involved!
#Gemma4 #AI #OpenSource #ML
Sabela retweeted
Gemma is hiring! 🚀🚀🚀
Please apply if you want to work on pre-training and/or designing models at the intersection of multiple modalities (text, audio, vision, video, actions,)
Location: Paris
Apply here: job-boards.greenhouse.io/dee…
Sabela retweeted
Our post-training team (once again) went beyond to post-craft this fantastic performance-for-size model:
@johanferret @ramealexandre @sarah_perrin_ @CdrGeo @nino_vieillard @sabelaraga @angelinepouget V Carbune L Rouillard L Hussenot @liu_gael @sinopalnikov @OlivierBachem
2/2
Sabela retweeted
Gemma 3n is out! 🚀🚀🚀
The frontier models from a year ago can now run locally on a phone!
Lots of innovations (e.g. matformers, mix’n’match, per layer embeddings) to make this model mobile first. And we finally have audio/video as an input for Gemma models!
1/2
Sabela retweeted
Releasing Gemma 3n, our new open-weight model processing audio, images and text (with improved multilingual capabilities), optimized for on-device usage with MatFormer architecture (enabling adaptive compute) and reaching 1283 on Chatbot Arena. Read more: developers.googleblog.com/en….






















