Play online math games in agent mode
The author had GPT-5 in ChatGPT agent mode play online versions of Math Games with Bad Drawings, and it mostly worked, unlike earlier rental car and tax form attempts.
Done withChatGPT Agent
- What they did
- The author pointed GPT-5 in ChatGPT agent mode at online versions of games from Ben Orlin's Math Games with Bad Drawings. The agent clicked around the game sites in its own browser and chose moves, with its reasoning visible. The author played the opposing side and had to type "your turn" after each move, since the agent couldn't tell when it was its turn.
- How it went
- Basically worked, which beat earlier attempts at rental car booking and tax forms. It played poorly, though. The author beat it easily as both Bob and Alice, and it wrongly claimed it could watch for turn changes.
- Worth knowing
- The agent can't watch a page for changes, so you have to prompt it every turn. A commenter also noted the games may be in its training data.
Try it yourself with ChatGPT Agent
Open [URL of an online game, e.g. a math or puzzle game] and play it for [number] rounds, trying to get the highest score you can. Tell me your score and describe where you struggled. Stop if the site asks for login, payment or personal information.
Source: agifriday.substack.com · Undated
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Plan a family trip and pre-fill bookings
A reporter had ChatGPT Agent plan a 4-day San Diego trip: it searched flights, compared hotels, built an itinerary and pre-filled booking forms, stopping before purchase.
Plan Saturday errands and book haircuts
ChatGPT Agent planned errands, found salons with online booking and worked on booking kids' haircuts, with some steps needing human confirmation.
Build a benchmark image collage in Canva
The author had ChatGPT Agent log into Canva and build a collage of benchmark results; after over 75 minutes the result was poor as it fumbled the interface and drew shapes in Python.
Play RollerCoaster Tycoon
Ramp Labs put Claude Code in RollerCoaster Tycoon.
- Claude
Let agents play computer games all day
Agents including Claude Opus and o3 were given computers to independently play games like 2048 for hours.
Build a match-3 game from a single prompt
An author prompted Gemini 3 Pro once to produce a working Candy Crush-style game with a shareable link, and says a non-technical friend did too.