@premsai-pendela· Jun 17
Debugged Naga's Agent
Naga's Agent
MARATHON
8.4x your typical session
600.6K tokens vs typical 71.8K
600.6ktokens
29prompts
4:10:11time
14lines
Codexgpt-5.5$16.81
mostly building
4h 10m session running commands and testing. 600K tokens across 29 prompts using gpt-5.5.
Reconstructed from session history
Tokens / prompt20.7k
Cost / line$1.201
Cache hit98%
Burn rate$4.03/hr
Derived from this session's token and cost data. Not shown on the feed.
Building 51%Thinking 49%
Comments