5.6 Terra high is underrated.
- Source
- x.com
- Date
5.6 Terra high is underrated. Switched @clawsweeper (GitHub review bot) to it and it's ~40% faster overall with negligible quality loss. Better than 5.5 on all counts. Massively cheaper. (Tried xhigh but that negates perf wins, didn't make a noticable difference in review evals)
Context
More reasoning effort did not improve one developer's code-review workflow. Peter Steinberger reports that switching his GitHub review bot to 5.6 Terra at high effort made it roughly 40% faster overall, with little quality loss in his .
When he tried extra-high effort, the speed benefit disappeared without a noticeable improvement in review evaluations. His account makes the effort setting part of the model-selection decision: the useful comparison was review quality and completion time together. The post does not supply the evaluation cases or detailed timing method.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
Checking sign-in…
Loading comments…