Users Criticize Opus 5 Performance: Frequent Apologies and Suspected Downgrade
PtrPomorski · x · 2026-08-15
A user criticized Anthropic's Opus 5 model, describing its performance as an absolute disaster. The user noted the model repeatedly outputting apologies and self-corrections like "my earlier claim was wrong," leading to a suspicion that it might be a downgrade similar to GPT-3.5 labeled as version 5.
More from Models
- Analysis suggests Qwen 3.8 27B distilled DeepSeek-R1 via GRPO — mishig25 · 2026-08-15
- Local LLM choice: Muse vs Qwen on 24GB VRAM for chat — Viktri1 · 2026-08-15
- Dev warns: GLM may bluff if it doesn't explicitly use 'analyze_image' — DevDminGod · 2026-08-15
- Qwen3.8 Review: Targeted Gains vs. General Regression — KitchenAmoeba4438 · 2026-08-15
- Alibaba's open-weight models top 3B downloads, beating Meta globally — kimmonismus · 2026-08-15
- Test of Gemini and Claude for melody generation fails — prajdabre · 2026-08-15