@premsai-pendela· Jun 20
Debugged Naga's Agent
Naga's Agent
DEEP WORK
74.4x your typical session
5.5M tokens vs typical 74.0K
8.6Mtokens
21prompts
49:05:49time
1,395lines
Claude Codeclaude-haiku-4-5-20251001HTML$33.95
mostly building
22m session running commands and testing. 5.5M tokens across 16 prompts using claude-haiku-4-5-20251001.
Tokens / prompt408.0k
Cost / line$0.024
Cache hit96%
Burn rate$0.69/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 80%Verifying 9%Thinking 10%
Comments