Gemma 3
Gemma 3 is Google's family of open-weight language models ranging from 1 billion to 27 billion parameters, with multimodal capabilities supporting image and text inputs, a 128,000-token context window, and support for over 140 languages. Recent deployments include NASA's in-orbit satellite imagery analysis aboard a spacecraft and integration into Google's privacy-focused data analysis system, while the company released Gemma Scope 2, an interpretability toolkit with sparse autoencoders to help AI safety researchers understand the model's internal workings.
Updated 3 August 2026
Specifications
No specifications recorded yet.
Latest developments
NASA Puts Google’s Gemma Large Language Model in Orbit
IEEE Spectrum AI · 1 week ago ·
8
Gemma Scope 2: helping the AI safety community deepen understanding of complex language model behavior
Google DeepMind · 7 months ago ·
5
Toward provably private insights into AI use
Google Research · 9 months ago ·
48
Vision Language Models (Better, faster, stronger)
Hugging Face Blog · 1 year ago ·
50
Mistral Small 3.1
Mistral AI · 1 year ago ·
32
Welcome Gemma 3: Google's all new multimodal, multilingual, long context open LLM
Hugging Face Blog · 1 year ago ·
26
2026
2025
Relationships
Products & technology
- Google develops this model · 4 sources
- Integrated with Hugging Face · 1 source
- Gemma Scope 2 integrated with this model · 1 source
- NASA deploys this model · 1 source
- Google deploys this model · 1 source