Pangram verdict · v3.3
We believe that this entire text is human-written.
2 %
AI likelihood · overall
Human 100% human-written
0% AI-generated
Analyzed
Aug 18
backend: pangram/v3.3
Segments scanned
1 windows
avg 83 words each
Distribution
100 / 0%
human / AI fraction
Verdict
Human
Pangram v3.3
Article text · 83 words · 1 segments analyzed
Human
AI-generated
§1 Human · 2%
Update We are investigating elevated errors on requests to Claude Mythos 5, Claude Fable 5, Claude Opus 5, Claude Sonnet 5, Claude Haiku 4.5, and other Claude models. We will provide an update as soon as possible. Posted Aug 18, 2026 - 16:20 UTC Investigating We are investigating reports of degraded performance affecting multiple models. We will provide an update as soon as possible. Posted Aug 18, 2026 - 16:20 UTC This incident affects: claude.ai, Claude API (api.anthropic.com), Claude Code, and Claude Cowork.