3 ms·
it may be the agent features in my case. now that i think about it, i also forgot that my CLAUDE.md is different from my AGENTS.md either way, all that one can
by merlindru 6mo ago
it may be the agent features in my case. now that i think about it, i also forgot that my CLAUDE.md is different from my AGENTS.md
either way, all that one can really rely on is the benchmarks, and those are easily cheated/overfitted to.
i think it's all very hard to quantify, so take my previous comment with a massive rock of salt