Qwen4-Max finds all 12 planted contradictions in a 199-page fake annual report
Div_pradeep · x · 2026-11-18
Testing Qwen4-Max's 1M-token context: the model worked through a fake 199-page annual report with 12 planted contradictions, cross-checked the CEO letter against notes buried 150 pages later, ranked the top 10 risks, and cited page numbers — work that would take a human analyst hours.
More from Models
- AI agents sometimes produce unreadable but correct math — burny_tech · 2026-11-18
- Code model eval: Rust hits 95-100% accuracy as 13 languages see big gains — shuding · 2026-11-18
- Asked Opus to make a video explaining consciousness and existence — "no wiser but cool" — flowersslop · 2026-11-18
- theo retests GPT-6.2 Astra: better, but still 'directionally very bad' at debugging — airesearch12 · 2026-11-18
- Claude correctly folds ubiquitin from a plain prompt, no specialized tools needed — iScienceLuvr · 2026-11-18
- GPT-6.2 Astra Tops Bug-Hunt Coding Benchmark at 62.7/105, One-Fifth Sonnet's Cost — PawelHuryn · 2026-11-18