Marathon debugging on backend
3.2M tokens vs typical 563.0K
3h 19m session running commands and testing. 3.1M tokens across 52 prompts using claude-opus-4-8.
Reconstructed from session history
Derived from this session's token and cost data. Not shown on the feed.
Comments