A society of 100 agents was built using the Jev model, enabling 100 decisions per request in approximately 250ms. This setup allows for 20,000 decisions at a cost of just 11 cents, making it 240 times cheaper than a frontier model.
added by @iamAliAgha
Ranked from stored criteria vectors. No live classification on this page.
Marionette was used to automate a Flutter game featuring five puzzles. It efficiently read the code and interacted with the game in 9.7 seconds at a cost of $0.0012.
This feed optimizer uses a single call to Jev to classify posts into categories like 'ragebait' and 'useful.' It scored 150 posts in under 10 seconds, achieving a 63% retention rate on unseen bookmarks.
A small MCP endpoint from OpenClaw was exposed over Tailscale and connected to Poke through its API, allowing the two agents to communicate privately and exchange requests and context.
Jev was tested in an AI customer feedback system using 1,040 GitHub issues, proving to be 98% cheaper and 84% faster than Claude Sonnet 4.6 while maintaining higher accuracy.
A space race game was developed for TypeSafe/Jev, achieving 31,382 points with 180 stars and 18 kills in the final 180 seconds. The replay video showcases the decision-making process and verified outcomes without new API calls.
The Jev model from Typesafe AI was tested on Modemdev's evaluation for agent replies, showing an 8x speed improvement and a 25x cost reduction.