@jaime-berkovich· 2h
Marathon debugging on automata_gpt_2_project
automata_gpt_2_project
DEEP WORK
Back after 14 days
Welcome back
4.7Mtokens
45prompts
22:03:18time
2,493lines
Claude Codeclaude-opus-5Python$165.16
mostly building
22h 3m session running commands and testing. 4.6M tokens across 45 prompts using claude-opus-5.
Tokens / prompt104.4k
Cost / line$0.066
Cache hit98%
Burn rate$7.49/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 66%Verifying 7%Thinking 27%
Comments