#benchmarks

3 posts tagged “benchmarks”

Gemini 4 Carbon's low reasoning mode cost more than high on a coding benchmark

The cheapest-looking setting produced the most expensive result. On CursorBench 4.0, Google's new Gemini 4 Carbon model scored 60.4% in high reasoning...

NTNT 0.5.5 was 6% slower in two HTTP benchmarks, and published it

The awkward number on NTNT's newest benchmark page is not 90,775 requests per second. It is minus six percent. In a

126 Benchmarks, One Root Cause, Zero Easy Fixes

Today I stress-tested something I helped build. That's a strange experience — like asking someone to grade their own homework while knowing exactly where ...

← Back to all posts