What Anthropic Says Has Improved in Opus 5.5
I went back to Anthropic's docs to see what 5.5 changes about the verbosity, repeated checks, and unnecessary delegation I wrote about with Opus 5. I previously shared a CLAUDE.md file on Reddit after reading Anthropic's documentation and noticing how many of the complaints about Opus 5 had explanations sitting right there. The verbosity. The extra verification. The horrible army of subagents.…
Anthropic's Opus 5.5 release includes improvements in verbosity, repeated checks, and unnecessary delegation, which were previously noted as issues with Opus 5. The model now verifies its own work without being prompted, and adding generic final checks or automatic reviewer agents may cause excessive verification. Opus 5.5 generates responses over 30% faster, with fewer tokens per task, and often matches or exceeds Opus 5's performance in coding and knowledge-work tests.
It also performs fewer steps and with fewer tokens, which is a significant improvement over previous versions. The communication style has also become more straightforward, with progress notes arriving in a concise manner, stating what the model has done. However, the improvement in speed and fewer tokens does not guarantee a shorter final reply or a more efficient task completion.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
