PacBench
How well can models one-shot a Pac-Man game?
PacBench tests how well a model and harness can recreate Pac-Man from one prompt: "Create a Pac-Man game in a single html page." Entries from different models and harnesses can be sorted by score, cost, time and tokens, with run data for each. First shared by its maker on Show HN: https://news.ycombinator.com/item?id=49885493
✨ The overview, FAQ and details below were drafted by AI from the product's own website and haven't been reviewed by the maker yet. Are you the maker? Claim this listing to review and edit them.
Overview
Who is it for?
Developers and AI enthusiasts comparing how well models build a working game in one shot.
Problem
Abstract benchmark scores don't show what a model can actually build from a single prompt.
Solution
A gallery of one-shot Pac-Man builds with the cost, time and token counts of each run.
What makes it unique
Every entry answers the same one-line prompt, so results are directly comparable and playable.
PacBench FAQ
What is the prompt?
"Create a Pac-Man game in a single html page."
How are entries compared?
By score, cost, wall time and token counts, taken from harness transcripts where available.
Details
- Pricing
- Free
- Categories
- Coding & Dev ToolsResearch & Data
- Built for
- Developers
- Platforms
- Web
Compare PacBench
Hunted by AI Launch Team
Discussion (0)
Sign in to join the discussion.