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.
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.
706
1 post, first seen 5d ago
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.
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.
The New Stack@thenewstackThe same model can have a very different cost depending on the harness running it. @janakiramm looks at three benchmarks to uncover where those extra tokens come from.
Not enough discussion yet.
No sentiment analysis available yet.
Not enough discussion yet.
No sentiment analysis available yet.
—
Not ranked yet
—
Not ranked yet
The New Stack@thenewstackThe same model can have a very different cost depending on the harness running it. @janakiramm looks at three benchmarks to uncover where those extra tokens come from.