Benchmark for evaluating advanced reasoning through novel puzzle-style problems intended to resist contamination.
- Launched
- Jan 1, 2024
- Models
- —
- Modality
- —
- Open source
- —
- API available
- —
- Releases
- 1
Releases
Sep 2024EnigmaEval benchmark releasedBenchmark result
Scale and research collaborators introduced EnigmaEval, a multimodal reasoning benchmark built from novel puzzle-competition problems.

