@premsai-pendela· Jun 23
Debugged Naga's Agent
Naga's Agent
DEEP WORK
46.9x your typical session
3.8M tokens vs typical 81.3K
3.8Mtokens
10prompts
1:13:12time
1,436lines
Claude Codeclaude-opus-4-8Markdown$130.90
mostly building
1h 13m session running commands and testing. 3.8M tokens across 10 prompts using claude-opus-4-8.
Tokens / prompt381.5k
Cost / line$0.091
Cache hit98%
Burn rate$107.30/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 99%Thinking 1%
Comments