Benchmark
BrokenArXiv Benchmark
What is BrokenArXiv?
arXiv proofs with planted errors; tests whether a model catches a broken argument instead of reproducing it. Benchmark Radar tracks 1 reported scores for…
- Evaluates
- math
- Openness
- unknown
- Importer
- Model reports
- Reported scores
- 1
Source provenance
- Original evidence https://matharena.ai/brokenarxiv
BrokenArXiv links
- Website https://matharena.ai/brokenarxiv
BrokenArXiv results and reported scores
Model reports
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Hy4 preview | Tencent | 54.6 | 2026-08-28 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.
Which sources cite BrokenArXiv?
- Hy4 preview Tencent, 2026-08-28