SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful
Why CEREBRO kept it
Coding-agent benchmark comparing Opus 5.5 and reasoning-effort settings.
The text below is an automated extraction of the article at https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/benchmarking-new-models-on-slop-code-bench.md, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (github.com).
SlopCodeBench update 10/2026 - Opus 5.5 update, Astra improves by 1.5% over gpt-5.5, xhigh considered harmful October 15, 2026 - Dex Horthy w/ contributions from Gabe Orlanski NOTE: This is Part I of our October 2026 writeup: the Q4 state of the slop. Gabe and the SlopCodeBench team will share their analysis and a more qualitative deep dive into slop patterns in Part II. For my previous writeups on SlopcodeBench, see benchmarking-opus-5-on-slop-code-bench.md and benchmarking-sol-fable-kimi-on-slop-code-bench.md. Both include deeper dives on the core structure of SCB and how each of its challen
Community take
Commenters say medium effort gives the best output for most tasks, and xhigh only pays off for spatial reasoning or creative writing, since higher effort mostly adds unrequested output.
Backlinks
Appeared in 1 briefing