Subvolts take Google's Nano Banana model pushes vision-language AI forward with advanced image editing and reasoning. Sep 23, 2025 - Developers
- Learn about Nano Banana's capabilities for image generation, editing, and integration with world knowledge for new applications.
- Business
- Understand how advanced vision-language models can expand creative and practical use cases for AI in your industry.
This episode provides a detailed look at a significant new multimodal model from a leading AI lab.
Inside Nano Banana 🍌 and the Future of Vision-Language Models with Oliver Wang - #748
Audio streams from Sam Charrington. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Sam Charrington interviews Oliver Wang, Google DeepMind principal scientist and tech lead for Gemini 2.5 Flash Image (Nano Banana). They discuss the shift to multimodal agents, Nano Banana's image generation and editing capabilities, its integration with Gemini's knowledge, and the future of vision-language models. The conversation covers model maturity, scaling, emergent behaviors, evaluation challenges, and risks of AI-generated training data.
This episode explores a new AI model called Nano Banana that can understand and create images. It's like an AI artist that can also use its world knowledge to make its creations more useful and creative. People care because this technology could change how we create and interact with digital content.
This episode offers insight into Google's latest vision-language model, Nano Banana, and its place in the evolving landscape of multimodal AI. It's for researchers, developers, and AI enthusiasts tracking frontier model development. The discussion on image generation, editing, and world knowledge integration connects directly to current efforts by Google Gemini and other labs like OpenAI and Anthropic to build more capable and versatile AI systems.
- HostSam Charrington
- GuestOliver Wangprincipal scientist, Google DeepMind
- Guest: Oliver Wang, principal scientist at Google DeepMind and tech lead for Gemini 2.5 Flash Image
- Title: Inside Nano Banana 🍌 and the Future of Vision-Language Models with Oliver Wang - #748
- Show: The TWIML AI Podcast
- Length: 64 minutes
- Model Code Name: Nano Banana
- Model Product Name: Gemini 2.5 Flash Image
Today, we’re joined by Oliver Wang, principal scientist at Google DeepMind and tech lead for Gemini 2.5 Flash Image—better known by its code name, “Nano Banana.” We dive into the development and capabilities of this newly released frontier vision-language model, beginning with the broader shift from specialized image generators to general-purpose multimodal agents that can use both visual and textual data for a variety of tasks. Oliver explains how Nano Banana can generate and iteratively edit images while maintaining consistency, and how its integration with Gemini’s world knowledge expands creative and practical use cases. We discuss the tension between aesthetics and accuracy, the relative maturity of image models compared to text-based LLMs, and scaling as a driver of progress. Oliver also shares surprising emergent behaviors, the challenges of evaluating vision-language models, and the risks of training on AI-generated data. Finally, we look ahead to interactive world models and VLMs that may one day “think” and “reason” in images.
The complete show notes for this episode can be found at https://twimlai.com/go/748.
- vision-language models
- Oliver Wang
- AI that understands and works with both images and text.
- multimodal agents
- Oliver Wang
- AI systems that can process and use different types of information, like text, images, and sound.
53 min1 h 30Why Jev Is Changing How We Build With AI with Diogo Almeida - #779
WithDiogo Almeida
1 h 081 h 05From Voice Agents to AI Avatars with Alexander Smola - #777
WithAlexander Smola
Same guests, other shows
41 minEp 75: Nano Banana’s Oliver Wang and Nicole Brichtova - Behind the Breakthrough as Gemini Tops the Charts
WithNicole Brichtova, Oliver Wang
28 minProtection, Provenance, and Prompts: YouTube’s Angela Nakalembe
WithAngela Nakalembe
1 h 03
1 h 08New top score on ARC-AGI-2-pub (29.4%) - Jeremy Berman
WithJeremy Berman
1 h 43The Latest in AI: Job Loss, Elon & Sam Altman Chip Race & the "AI Bubble" w/ Brian Elliott (Blitzy), Emad Mostaque & Dave Blundin | EP #197
WithBrian Elliott, Emad Mostaque, Dave Blundin
Questions people ask
- What is Nano Banana?
- Nano Banana is the code name for Gemini 2.5 Flash Image, a frontier vision-language model developed by Google DeepMind that can generate and edit images.
- What are the challenges in evaluating vision-language models?
- The episode discusses the difficulties in assessing the performance and capabilities of vision-language models, including potential risks associated with training on AI-generated data.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Sam Charrington.