Developer kalomaze calls Opus 5 a bad model lacking self-awareness

Prominent open-source developer kalomaze called Anthropic's Opus 5 a bad model, citing its tendency to state assumptions as proven facts and its lack of self-awareness about its own weaknesses, where Sonnet performs better.

2026-09-13 ~ 2026-09-13 · 3 related posts

• kalomaze: Opus 5 is "such a bad model" — kalomaze · 2026-09-13
• Smaller AI models assert hypotheses as proven facts and ignore evidence already in tool calls — kalomaze · 2026-09-13
• Dev: Opus 5 lacks awareness of its own weaknesses, unlike Sonnet — amplifiedamp · 2026-09-13