Open source alternatives to Pac-Bench
200 open source apps like Pac-Bench, closest match first.
Top picks
- A

AI Gallery
LaunchThis site is a collection of open source classic games and puzzles written in React that you can play for free in the browser. It includes Minesweeper with custom board size and mine count, a sliding number puzzle that can show a solution,
FreeOpen source
- B

Benchmark Heaven
LaunchBenchmark Heaven compares AI models on cost and capability, with benchmark results, charts and provider prices. It is the home of JevBench, a decision-model benchmark that measures accuracy, calibration, latency and cost. The site has a cos
Free
- G
GameBlocks
LaunchConcise, self-explanatory building blocks for AI coding agents to prototype browser-based 3D games.
FreeOpen source
- N
Noobi.ai
LaunchNoobi.ai is a local desktop agent that turns a game idea into a playable browser game. You describe the game in a prompt, and the agent chooses a Web project or a Godot 4 project and builds it inside an isolated project folder. The work run
FreeOpen source
- P

Pixtuoid
Launchpixtuoid is a terminal app that draws a pixel-art office for your AI coding agents. Each coding session appears as a small animated coworker that walks in, takes a desk and gets to work, shown with typing, wandering and coffee runs. The off
FreeOpen source
- P

Petdex
LaunchPetdex is the public gallery of animated pets for coding agents. Works with ChatGPT (Codex), Claude Code, Gemini CLI, OpenCode, Qoder, Kimi Code, CodeBuddy, and OMP. Browse open-source pets, preview their animations, and install one with a
FreeOpen source
More like Pac-Bench
About Pac-Bench
Pac-Bench
PacBench: how well a model and harness can recreate Pac-Man from a single prompt
PacBench is a web gallery that compares how well different AI models and coding harnesses can build a Pac-Man game from one short prompt: create a Pac-Man game in a single HTML page. Each entry is shown as a card that you can open and play, and the gallery can be searched, filtered by model, and sorted by score, cost, time, tokens or release date. Where the harness transcripts allow, cards record wall time, token counts, cost and the size of the generated HTML file. The notes explain how runs were made, for example that some phases used Claude Code pointed at OpenRouter, others used Antigravity with Gemini, and how costs were calculated for particular entries. It is a useful reference for developers curious how coding agents compare on a small, self-contained game task, and the source is on GitHub.
- Category
- AI
- Website
- jonclegg.github.io

