@jeremylongshore· 3h
Debugged diagnostic-pro
diagnostic-pro
DEEP WORK
7.6x your typical session
16.7M tokens vs typical 2.2M
16.7Mtokens
24prompts
142:20:45time
2,343lines
Claude Codeclaude-opus-5Python$256.10
mostly building
142h 20m session running commands and testing. 16.6M tokens across 24 prompts using claude-opus-5.
Tokens / prompt695.5k
Cost / line$0.109
Cache hit94%
Burn rate$1.80/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 55%Verifying 17%Thinking 11%Delegating 18%
Comments