MicroEvals
Public evaluations
Showing 261-280 of 6655
12 prompts sampling FSC architecture, engineering, discovery, security, and migration scenarios for model comparison.
A set of difficult tasks with the theme of emoji's including: tier list creation, music creation and, themeable website creation.
make a single html in which the world globe is there (do not svg map) and the map is highlighted in a glow while line all around borders of countries and it will be in dark mode and make the globe 3d and good looking and interactive with satellites roaming around with dotted path make sure it works . let me be very clear 1. I want full black earth, land will be in a very dark grey colour , and ocean must be black. The add a white outline to every country borders. 2. Try to make code as much ad low bugs possible. Add a feature when we tap any country it open slides up a info card containing all info about that country. Keep ui of that in transulent black. And cover all countries
How well can AI create fun, little games as web apps that can be played on mobile devices? This benchmark tests AI’s abilities to generate good mobile UI and controls as well as basic gameplay experiences.
local test
Which makes the best website
Compare multiple AI models on accuracy, reasoning, coding, mathematics, and general problem-solving ability using the same prompts.
Cardiac phisiology quiz.
A five-scenario benchmark designed to evaluate an LLM’s ability to perform as a senior Cloud Architect, DevOps Engineer, and SRE. The test covers production Kubernetes architecture, incident response, cloud security, legacy-to-cloud migration, and FinOps. It focuses on technical correctness, practical decision-making, trade-off reasoning, scalability, reliability, and depth of Cloud and Kubernetes expertise.
