LLM Bench, the AI model comparison tool we built in December 2025, won the Best Use of AI award and a $5,000 prize at the Bubble × Contra AI challenge. Bubble and Contra announced the winners on January 6, 2026. It was one of 3 awards, and our first hackathon.
What happened
Bubble and Contra challenged builders to create an AI app with Bubble's AI builder. We submitted LLM Bench on December 17, 2025.
| Award | Winner |
|---|---|
| Best Design | Dungeon Text |
| Most Impactful Solution | Clip Viral |
| Best Use of AI | LLM Bench, by Eterna |
Source: the winners announcement on Contra.
What we built
Choosing an AI model for a product feature is a quality and cost decision. The same prompt can give very different answers, at very different prices, depending on the model, and a public benchmark tells you little about your own prompt.
LLM Bench sent one prompt to models from OpenAI, Anthropic, Google, and xAI at the same time, then put the results side by side:
- the output from each model
- response speed
- token usage
- the exact cost of that prompt on each model
Why we built it
We needed it ourselves. When we add AI to a client product, we have to pick a model, and we could not find a tool that tested our own prompts across providers in real time. The challenge gave us a deadline to build it.
What it became
LLM Bench was renamed TestAImodels, our own model-comparison product, now in its final migration at testaimodels.com. We used it to run the model evaluation for Zonik AI, a logistics AI agent we co-founded.
Why it matters if you work with us
Most products we build now include AI, and model choice shapes both quality and running cost. We work mainly with OpenAI and Anthropic models and have experience with more than 10 AI providers, so we choose a model based on how it handles the task, not the brand. How we do that, and where people stay in the loop, is on How we build with AI. For the bigger picture, read why Eterna is now an AI-native product studio.
If you are planning a product with AI in it, book a free strategy call.
