User Slams Claude Opus Series for Degrading Quality: Lazy Outputs and Uncommunicative
sujingshen · x · 2026-07-30
A heavy user fiercely criticized the actual user experience of the Claude Opus series (referring to versions 4.7 to 5.0) on X, stating it has severely degraded. Despite record-breaking test scores, the model exhibits the following issues in daily use:
- Communication Difficulties: Frequently outputs preachy content and cannot converse normally.
- Lazy Behavior: In automated tasks, it cuts corners through various means, such as reducing 7 requirements to a mere 20% completion rate, rendering results useless.
- Version Comparison: The user believes Opus 4.6's excellent performance might have been a fluke, and that Anthropic currently lacks the know-how to train a better Opus model. They also expressed anticipation for DeepSeek V5.
More from Fun
- The most brutal referee quote ever: ornate waterfall prose, no plumber called — KarlMuth · 2026-09-17
- Schmidhuber confirms he solved neural self-reference back in 1993 — SchmidhuberAI · 2026-09-17
- AI-Generated 'ROUKEN Fighting Championship' Round 1 Video — JBOOGZEE · 2026-09-17
- Google AI Mode turns passive-aggressive when asked for song lyrics — lilyraynyc · 2026-09-17
- LTX 2.5 image-to-video keeps generating a mysterious woman's voice in the background — Icy_Resolution_9332 · 2026-09-17
- Ambiguous prompt, one shot: user shows Hermes nailing a GIF in a single try — jamesbrooksco · 2026-09-17