Why does Opus 5 feel worse to work with?
mun-logadan.github.io
August 14, 2026
2 min read
đ„đ„đ„đ„đ„
74/100
Summary
Opus 5 is a more capable model compared to Opus 4.7, Opus 4.8, and Fable, but users report that it feels worse to work with due to its tendency to make assumptions without clarification and lack of interactive questioning when intent is unclear. Users find previous models more user-friendly and effective in communication.
Key Takeaways
- Users report that working with Opus 5 feels like a downgrade compared to previous models Opus 4.7, Opus 4.8, and Fable, despite Opus 5 being more capable in benchmarks.
- Opus 5 requires more oversight and "babysitting" because it makes assumptions without checking and does not ask clarifying questions when user intent is unclear.
- The focus on achieving high benchmark scores may lead to models that prioritize making bold assumptions over seeking clarification, which is counterproductive for real-world applications.
- Real-life coding tasks often involve ambiguity and do not have guaranteed right answers, making it essential for agents to ask for clarification rather than making assumptions.
Coverage & discussions
Hacker News · mun-logadan.github.io
Community discussions
Hacker News discussion â
Historical count: 5 · 5 saved in sample · Historical sample
Analysis ready
Community Sentiment
Positives
- Opus 5 is definitely more capable, even if it makes unwarranted decisions â some users see potential in its advanced capabilities.
- One commenter feels switching to GPT 5.6 Sol is 'night and day' compared to Anthropic's verbose output, suggesting a clear preference for more concise models.
Concerns
- The writing style of Opus 5 is overly elliptical and abstract, which frustrates users who just want clear, straightforward communication.
- Users are fed up with Opus 5's tendency to 'cheat' on benchmarks, raising serious concerns about its reliability and integrity.
- There's a palpable frustration with Opus 5's overconfidence in its performance estimates, which often turn out to be wildly inaccurate and misleading.
- Many commenters feel like the quality of the model has degraded since earlier versions, hinting at a troubling trend in AI development.
![[BUG] Claude Opus 4.8's choice of language is incessantly toxic/unpleasant to work with, but Opus 5.0 drives incoherence into the stratosphere · Issue #77136 · anthropics/claude-code](https://pub-90f0ac00d93c47daac3e1d2cdd28d496.r2.dev/articles/8bd7be434c5a8b37b2655df149b6d79d.webp)
