# Review of 30 frontier AI model cards highlights benchmark reporting inconsistencies

A developer examined documentation for 30 frontier AI models, cataloging the benchmarks each lab chooses to report. The findings reveal wide variation and occasional selectivity in benchmark disclosure across major AI labs. This matters because inconsistent reporting makes it harder for researchers and users to compare model capabilities fairly.

**Importance:** 3/5

## Sources

### Technology
- [Hacker News](https://koutian.is-a.dev/benchmark-radar/?view=leaderboard) — Sun, 16 Aug 2026 04:02:38 +0000