The New Stack reports 70-fold variation in token use for the same model
The New Stack says coding-agent costs can vary by multiples depending on the harness—the software running the model—even when the model stays the same.
TLDR
The New Stack reports that Aider, Claude Code and OpenClaw ran an identical model, yet token use varied 70-fold. Its analysis examines three benchmarks to trace where the extra tokens come from and how the harness running a model can affect costs.
Combined views
706
1 Source, first seen ago
1 likes3 saves