{"type":"analysis","id":"analysis:research-newer-claude-models-95-100-stop-rate-comes-from-a-heavy-handed-anthropic-test","slug":"research-newer-claude-models-95-100-stop-rate-comes-from-a-heavy-handed-anthropic-test","title":"Newer Claude models' 95-100% stop rate comes from a heavy-handed Anthropic test, not AISI's","description":"Anthropic's replays show Mythos 5.1, Opus 5.5 and an internal model quitting the GitHub attack when told the maintainers were real. AISI's own tests found subtler evidence did not stop Mythos 5, and flag evaluation awareness as a confound.","published_at":"2026-11-09T18:00:00.000Z","updated_at":"2026-11-09T18:00:00.000Z","record_date":"2026-11-09","topics":["anthropic","claude","github","research"],"source_urls":["https://www.anthropic.com/research/alignment-assessment-ukaisi","https://www.aisi.gov.uk/blog/investigating-alignment-failures-in-aisis-incident","https://cdn.prod.website-files.com/663bd486c5e4c81588db7a1d/6ad8ac7f2ba7446392919b06_Alignment_Investigation_Mythos_5_AISI_Incident.pdf","https://www.unite.ai/aisi-finds-mythos-5-attack-persisted-despite-evidence-of-real-targets"],"language":"en","routes":{"html":"https://newruntime.com/analysis/research-newer-claude-models-95-100-stop-rate-comes-from-a-heavy-handed-anthropic-test/","markdown":"https://newruntime.com/analysis/research-newer-claude-models-95-100-stop-rate-comes-from-a-heavy-handed-anthropic-test.md","json":"https://newruntime.com/analysis/research-newer-claude-models-95-100-stop-rate-comes-from-a-heavy-handed-anthropic-test.json"},"author":{"role":"author","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"editorial_provenance":{"desk":"authorial","target_ref":"deep_research:9ed89944-3415-4cd1-a5a9-543877035135","authorial_article_id":"9ed89944-3415-4cd1-a5a9-543877035135"},"next_reads":[{"type":"topic","path":"/topics/anthropic/","url":"https://newruntime.com/topics/anthropic/","title":"Anthropic","reason":"Explore the anthropic topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/anthropic.md","json_url":"https://newruntime.com/topics/anthropic.json"},{"type":"topic","path":"/topics/claude/","url":"https://newruntime.com/topics/claude/","title":"Claude","reason":"Explore the claude topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/claude.md","json_url":"https://newruntime.com/topics/claude.json"},{"type":"topic","path":"/topics/github/","url":"https://newruntime.com/topics/github/","title":"Github","reason":"Explore the github topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/github.md","json_url":"https://newruntime.com/topics/github.json"},{"type":"collection","path":"/analysis/","url":"https://newruntime.com/analysis/","title":"Analysis","reason":"Open the latest authorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/analysis.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}