Shipped v1.1.0 with pricing and parallelism fixes
3.1M tokens vs typical 22.1K
Fixed cost calculation and session parallelism metrics to eliminate accuracy bugs.
Derived from this session's token and cost data. Not shown on the feed.