A user reported that Anthropic's Claude Opus 5 (High) model exhibited significant hallucination issues while assisting with API service catalog automation. When asked for the next steps to complete a feature, the model incorrectly suggested merging three pull requests, deviating from the original plan of one pull request plus one for testing. The user also noted that the model provided a nonsensical response when questioned about the identity of the third pull request. AI
IMPACT Highlights potential reliability issues in advanced models, impacting user trust and adoption for complex automation tasks.
RANK_REASON User report of model hallucination, not an official release or benchmark.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →