←── back to feed
/topics/claude-opus-5-anti-verbosity-policy

Claude Opus 5 anti-verbosity policy

2 items1 sourcesupdated 32d agotrend 0

Anthropic introduced an anti-verbosity policy for Claude Opus 5 that constrains response length and complexity. The model achieved 100% on the ARC-AGI-3 public benchmark when paired with Claude Code and an additional skill.

  • Claude Opus 5 implements anti-verbosity constraints to reduce unnecessary response length
  • Model reached 100% accuracy on ARC-AGI-3 public benchmark
  • Achievement required Claude Code integration plus one additional skill
  • Policy aims to improve response efficiency and directness
  • Benchmark result represents full performance on public ARC-AGI-3 test set