Vibing with the flow, Marketing @kombaico - Your AI design engineer
Joined April 2025
- Tweets217
- Following116
- Followers50
- Likes601
Pinned Tweet
Bookmark this and thank me later.
kombai.com/gallery
(20000+ UIs for agents & humans - free forever, no signup required)
8
342
Free & useful resources for design engineers 👇
1. designeer.xyz → tools & inspo
2. bencho.dev → interactive React components
3. inspora.design → daily UI inspo
4. designengineer.tools → tools & agents
5. dialkit.dev → real-time UI tuning
6. kombai.com/gallery → 20K+ UIs for agents
Any others?
6
227
👇 Free go-to brand design resources 👇
20K+ UIs for agents → kombai.com/gallery
UI Interactions → bencho.dev
Decks → deck.gallery
Launch videos → launch.gallery
Inspirations → recent.design
Logos → logosystem.co
Tools & Resources → designeer.xyz
Any others?
1
6
764
Must-have tools for design engineers -
- kombai.com - Your AI design engineer
- quiver.ai - Generate, edit, animate vector graphics
- 21st.dev - Component library for agents
- designeer.xyz - Curated resources for designers
- recent.design - Design inspirations from the web
Save it; you'll need it later.
5
234
If you're designing with AI, these are must-haves -
1. kombai.com/gallery - UI inspirations for you & agents
2. recent.design - curated top-notch designs
3. bencho.dev - Interactive React components
4. posts.design - Social post inspirations
5. Designeer.xyz - Resources for design engineers
Save it; you'll need it later.
2
7
399
The best collection of curated designs from the internet.
Checkout - kombai.com/selects
5
195
Sourabh Upreti retweeted
Gemini 4 Argon by @GoogleDeepMind has landed on Design Creativity Bench!
Argon is one of the top-performing models on Design Creativity Bench, combining strong creative performance with improved adherence to design requirements.
It comes in at #6 overall with a Creativity Score of 55.7 ahead of Claude Opus 5.5 and Fable 5.1.
It also shows strong performance in following design requirements, achieving 98.6% appropriateness, up from 97.5% for Gemini 3.8 Flash (up 1.1%).
Congrats to the @GoogleDeepMind team on the release!
Sourabh Upreti retweeted
Most frontend and design benchmarks evaluate whether models produce working UIs and/or pairwise design preferences of judges.
With this benchmark, we measure something different: the diversity of design decisions a model makes when the problem leaves room for creative exploration.
We construct paired briefs that preserve the underlying UI task while changing the product context.
For example, a clinical dashboard and a pharmacy operations dashboard share related functional requirements but call for different information architectures and visual hierarchies.
Models receive no prescribed layout or visual direction.
We then compare their generated interfaces using a similarity metric that measures structural composition and information architecture, with color evaluated separately.
This allows us to measure whether a model meaningfully changes its design approach, rather than merely changing content or styling.
⚠️ Frontier AI models still fall significantly short of human designers on design creativity. ⚠️
Introducing Design Creativity Bench, a new benchmark evaluating AI models’ creativity on UI design tasks.
Despite rapid advances in their ability to generate functional websites and product UIs, leading AI models still struggle to make creative design decisions.
We evaluated 16 frontier models across 80 open-ended UI tasks, each adapted to two distinct product domains alongside human designer references.
Existing UI benchmarks often measure a single output quality metric. Instead, we measure three dimensions of design creativity:
- Originality: How distinct a model's designs are from other models' outputs for the same brief.
- Creative Range: How much a model adapts its design decisions when the product domain changes.
- Appropriateness: How well the generated UI satisfies the functional requirements of the brief.
The results reveal a substantial gap in overall creativity: The best model, @OpenAI GPT-6 Astra, scores 65.5 against a human reference of 92.3.
We're publishing the full leaderboard, methodology, and reproducible evaluation framework to make design creativity a measurable dimension of frontier model capability.
📊 Benchmark Leaderboard: kombai.com/benchmark/
📄 Paper: arxiv.org/abs/2610.11539
🤗 HuggingFace Datasets: huggingface.co/datasets/Komb…
🔗 Github: github.com/kombai-io/design-…
Sourabh Upreti retweeted
⚠️ Frontier AI models still fall significantly short of human designers on design creativity. ⚠️
Introducing Design Creativity Bench, a new benchmark evaluating AI models’ creativity on UI design tasks.
Despite rapid advances in their ability to generate functional websites and product UIs, leading AI models still struggle to make creative design decisions.
We evaluated 16 frontier models across 80 open-ended UI tasks, each adapted to two distinct product domains alongside human designer references.
Existing UI benchmarks often measure a single output quality metric. Instead, we measure three dimensions of design creativity:
- Originality: How distinct a model's designs are from other models' outputs for the same brief.
- Creative Range: How much a model adapts its design decisions when the product domain changes.
- Appropriateness: How well the generated UI satisfies the functional requirements of the brief.
The results reveal a substantial gap in overall creativity: The best model, @OpenAI GPT-6 Astra, scores 65.5 against a human reference of 92.3.
We're publishing the full leaderboard, methodology, and reproducible evaluation framework to make design creativity a measurable dimension of frontier model capability.
📊 Benchmark Leaderboard: kombai.com/benchmark/
📄 Paper: arxiv.org/abs/2610.11539
🤗 HuggingFace Datasets: huggingface.co/datasets/Komb…
🔗 Github: github.com/kombai-io/design-…
Sourabh Upreti retweeted
Humans vs AI - who's more creative?
Looks like we finally have the answer 😄
⚠️ Frontier AI models still fall significantly short of human designers on design creativity. ⚠️
Introducing Design Creativity Bench, a new benchmark evaluating AI models’ creativity on UI design tasks.
Despite rapid advances in their ability to generate functional websites and product UIs, leading AI models still struggle to make creative design decisions.
We evaluated 16 frontier models across 80 open-ended UI tasks, each adapted to two distinct product domains alongside human designer references.
Existing UI benchmarks often measure a single output quality metric. Instead, we measure three dimensions of design creativity:
- Originality: How distinct a model's designs are from other models' outputs for the same brief.
- Creative Range: How much a model adapts its design decisions when the product domain changes.
- Appropriateness: How well the generated UI satisfies the functional requirements of the brief.
The results reveal a substantial gap in overall creativity: The best model, @OpenAI GPT-6 Astra, scores 65.5 against a human reference of 92.3.
We're publishing the full leaderboard, methodology, and reproducible evaluation framework to make design creativity a measurable dimension of frontier model capability.
📊 Benchmark Leaderboard: kombai.com/benchmark/
📄 Paper: arxiv.org/abs/2610.11539
🤗 HuggingFace Datasets: huggingface.co/datasets/Komb…
🔗 Github: github.com/kombai-io/design-…
Sourabh Upreti retweeted
Has AI replaced human designers as your source of design inspiration?
Or are you still turning to places like Dribbble, Behance, and designer portfolios for ideas?
We're launching a benchmark on AI design creativity next week, and we're curious how your creative process has changed.
Designers and builders, where are you finding inspiration these days? Tell us in the replies.