Battleships ranks 366 agent runtimes

Battleships is a useful antidote to glossy sandbox marketing, but its rankings still deserve scrutiny because the project comes from boat.dev, a competing provider.

• –Models monthly costs around realistic agent workloads instead of headline hourly rates
• –Compares practical primitives such as persistence, snapshots, SSH, Docker, networking, and concurrency
• –Publishes machine-readable data and source evidence, making the directory useful to coding agents and developers
• –Open-source updates and pull requests create a path for providers and users to correct stale pricing or feature claims
• –Results should be validated for region, security boundaries, egress, and workload-specific performance before switching providers