Google unveils Gemini Nano Banana 2.1 and EmbeddingGemma 2 AI models: Check key updates
Follow UsGoogle has introduced two new AI models, Gemini Nano Banana 2.1 and EmbeddingGemma 2.
Google describes Gemini Nano Banana 2.1 as a "high-efficiency image generation and conversational editing model."
Google describes EmbeddingGemma 2 as "the most capable model for on-device multimodal embeddings."
Google has introduced two new AI models, Gemini Nano Banana 2.1 and EmbeddingGemma 2. The first is focused on better image generation. The second is a multimodal embedding model that can work with text, code, images, video and audio. Nano Banana 2.1 is now rolling out across several Google products, including the Gemini app, AI Mode in Search and Google AI Studio. Meanwhile, EmbeddingGemma 2 is aimed mainly at developers.
SurveyGemini Nano Banana 2.1
Gemini Nano Banana 2.1 is an updated version of Nano Banana 2, also known as Gemini 3.1 Flash Image. Google describes it as a “high-efficiency image generation and conversational editing model.”
Also read: Apple partners with LG to launch smart doorbell, lock and other home devices: Report
Meet Nano Banana 2.1, our latest image generation and editing model 🍌 This upgraded version outperforms our previous models across the board, with notable leaps in visual design, mask-based editing, and subject consistency to help you create more natural-looking images.
— Google (@Google) October 6, 2026
The model is designed to create more realistic images and follow user prompts more accurately. It supports 1K, 2K and 4K image generation. Google has also fixed tiling issues that could appear when creating very wide or panoramic images.
Text inside generated images should now be clearer. Google has also improved infographic layouts. The new model also offers multi-image fusion, supporting up to 14 reference images. This can help maintain the appearance of characters and objects across different images. The model can keep up to four characters consistent and preserve details of up to 10 objects.
Nano Banana 2.1 is rolling out across the Gemini app, AI Mode in Search, Google AI Studio, Flow, Stitch, Google Ads and Gemini Enterprise Platform.
Also read: Apple’s October 2026 launch roadmap leaked: Smart home devices to OLED iPad mini, what to expect
EmbeddingGemma 2
Google describes EmbeddingGemma 2 as “the most capable model for on-device multimodal embeddings, natively mapping combinations of text, images, audio, and video into a unified embedding space.”
The model has 740 million parameters and is built on the Gemma 4 architecture. One of its main advantages is its focus on on-device use. “It can help find a specific video clip from a voice memo, or search through hours of audio recordings based on a text query, all processed by a single, natively multimodal model,” the tech giant said.
EmbeddingGemma 2 also has an 8K-token context window. Google claims this is four times larger than the previous version.
The new EmbeddingGemma 2 AI model is available through platforms including Hugging Face, Kaggle and more.
