Zenodo researchers document Claude Opus 4.6 returning zero-byte responses in 900 of 900 test cases

Researchers at Zenodo published a cross-vendor evaluation documenting Claude Opus 4.6 returning successful provider responses with zero visible UTF-8 output bytes in 900 of 900 test cases, distinct from explicit refusals or safety blocks. The study examined 31,430 total trials across 11 model identifiers from OpenAI, Anthropic, Google, and Moonshot, observing 11,658 void responses (37.09%) overall. The report preserves raw attempts, request configurations, and cryptographically linked event records for reproducibility.

Topics

Claude

Sources

This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.