Speculation: Opus 5 verbosity linked to watermarking detection
cool_architect · reddit · 2026-08-23
Users noticed Opus 5 has become noticeably more verbose and speculate it might be related to the new watermarking/SynthID feature. Statistical watermarking likely requires more tokens for reliable detection, so increasing response length could be a strategy to ensure consistent signals.
Related event: Users Suspect Opus 5 Verbosity Linked to Watermarking(2 posts)→
More from Models
- Ox Model Review: Strong One-Shot Performance, Struggles with Long Context — rachittshah · 2026-08-23
- Grok 4.6 achieves faster, cheaper tasks using fewer steps and tokens — XFreeze · 2026-08-23
- ProgramBench Results: Opus 5 Leads, 0813 Emerges as Strongest Open Model — teortaxesTex · 2026-08-23
- Qwen3.8-27B One Week Later: Community Verdict & Benchmarks — Jonathan_Rivera · 2026-08-23
- Fine-tuned Gemma 4 12B for 2.7x Better Tool Calling on 16GB VRAM — TheOneWhoWil · 2026-08-23
- Boson AI targets voice market with Higgs RealTime model supporting 100+ languages — smolix · 2026-08-23