Gemma 3n
Gemma 3n is Google's open mobile-first multimodal model, built to run entirely on phones, tablets and laptops. It accepts text, images, audio and video and returns text, ships in E2B and E4B sizes whose MatFormer and per-layer embedding design cut the working memory to roughly 2GB and 3GB, supports 140 languages of text and 35 for multimodal input, and offers a 32K context window.
- Type
- Multimodal
- Weights
- Closed
More from Google DeepMind
14
-
21 DL
AlphaFold 3
Text
AlphaFold 3 is Google DeepMind and Isomorphic Labs’ biomolecular structure model for predicting proteins, DNA…
-
2 DL
Chinchilla
Text
Chinchilla is DeepMind’s 70B language model used to demonstrate compute-optimal scaling with more training da…
-
25 DL
CodeGemma
Text
CodeGemma is Google’s open-weight Gemma coding model family for code completion, code generation, instruction…
-
Dolphin Gemma
Audio
DolphinGemma is Google’s audio model for learning patterns in dolphin vocalizations and generating dolphin-li…
-
Gemini
Text
Gemini is Google DeepMind’s closed-source natively multimodal AI model family for text, image, audio, video,…
-
Gemini 1.5
Text
Gemini 1.5 is Google’s multimodal Gemini generation focused on long-context understanding, with Gemini 1.5 Pr…
-
Gemini 2.0
Text
Gemini 2.0 is Google’s agentic Gemini generation, built for multimodal input, native tool use, and action-ori…
-
13 DL
Gemini 2.0 Flash
Text
Gemini 2.0 Flash is Google’s fast, multimodal Gemini 2.0 model for low-latency general tasks, tool use, and 1…
About Google DeepMind
1
AI research lab within Google developing frontier models, scientific AI systems, reinforcement learning methods, and products that support Google services and research.