Gemini 3.5 Flash now available in ai.to.design

Google has just launched a new model: Gemini 3.5 Flash. And we’re excited to announce that ai.to.design already supports it!

Gemini 3.5 Flash comes to ai.to.design

Gemini 3.5 Flash is Google’s latest model — and according to Google DeepMind, it delivers intelligence that rivals large flagship models at Flash-series speeds. It’s designed for agentic workflows, advanced reasoning, and multimodal tasks, with a 1 million token context window and support for text, image, video, audio, and PDF inputs.

The model’s high multimodal scores make it especially interesting for design generation in ai.to.design, as it translates directly to better spatial reasoning when generating UI layouts, handling visual references, and interpreting complex design briefs.

We put Gemini 3.5 Flash through its paces in ai.to.design, comparing it against Gemini 3 Flash with the following prompt:

Design the homepage of a meal-kit delivery service targeting home cooks. Include a hero section with a strong value proposition, a section showcasing this week’s featured recipes, a step-by-step “how it works” section, and a pricing plans comparison. Use a fresh, appetizing visual style with a light theme, clean layout and warm food photography placeholders.

Check out the results:

Analyzing the results

Gemini 3.5 Flash delivers a much higher quality result than its predecessor, not only closely matching the original brief, but also going above and beyond with additional sections and details that you would expect of a finished product.

• Fully on-brief: Our prompt requested a fresh visual style and warm phtotography, which the previous model didn’t quite deliver, as important visuals like the hero image were missing. Gemini 3.5 Flash perfectly delivered on the brief and exactly matched our theme requirements.

• Attention to detail: The result from Gemini 3.5 Flash offers superior attention to detail; with banners, labels and imagery that elevate the landing.

• Added extras: Gemini 3 Flash delivered the four sections requested in our prompt, however Gemini 3.5 Flash added extra sections that, while not requested in the prompt, made our landing more professional.

• Lightning fast: With the level of detail in the result delivered by the new Gemini 3.5 Flash, you would expect a much longer generation time than the previous model. However, it took just 15 seconds longer (0:57 vs. 0:42 seconds) to generate a much more complete design than Gemini 3 Flash.

Try it out!

You can now use ai.to.design to generate designs from prompts using Google’s latest model.

Gemini 3.5 Flash is live and available in ai.to.design. Give it a try!