@premsai-pendela· Jun 10
Marathon debugging on AI
AI
DEEP WORK
First 5M-token session
10.4M tokens
10.4Mtokens
266prompts
92:36:26time
4,725lines
Claude Codeclaude-fable-5Python$535.99
mostly building
92h 36m session running commands and testing. 10.4M tokens across 266 prompts using claude-fable-5.
Reconstructed from session history
Tokens / prompt39.3k
Cost / line$0.113
Cache hit98%
Burn rate$5.79/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 100%