Villanova2-2B-Base: benchmark results

Provider: VillanovaAI. Access: Open.

Aggregate Rating 1346 ± 32, rank #7797 of 10552 rated models, from 26 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Lara Mini Base Report (base-model protocol) - XNLI, All Languages (0-shot)41Accuracy (%) averaged over all languages/variants of the ben70.4
Lara Mini Base Report (base-model protocol) - ARC-Challenge (25-shot)55Accuracy normalized by answer length (%), English benchmark53.7
Lara Mini Base Report (base-model protocol) - Belebele, All Languages (0-shot)33Accuracy (%) averaged over all languages/variants of the ben50
Lara Mini Base Report (base-model protocol) - CRUXEval-I (2-shot)32.3pass@1, direct prediction, official answer parser (%)50
Lara Mini Base Report (base-model protocol) - HellaSwag (10-shot)70.4Accuracy normalized by answer length (%), English benchmark50
Lara Mini Base Report (base-model protocol) - Global PIQA (non-parallel), All Languages (0-shot)54.9Accuracy (%) averaged over all languages/variants of the ben46.3
Lara Mini Base Report (base-model protocol) - Belebele (English) (0-shot)63.9Accuracy (%), English benchmark44.4
Lara Mini Base Report (base-model protocol) - XNLI (English) (0-shot)52.9Accuracy (%), English benchmark44.4
Lara Mini Base Report (base-model protocol) - INCLUDE, All Languages (5-shot)38Accuracy (%) averaged over all languages/variants of the ben42.6
Lara Mini Base Report (base-model protocol) - T-HellaSwag, All Languages (0-shot)36.1Accuracy normalized by answer length (%) averaged over all l42.6
Lara Mini Base Report (base-model protocol) - Global MMLU, All Languages (5-shot)40.2Accuracy (%) averaged over all languages/variants of the ben40.7
Lara Mini Base Report (base-model protocol) - MATH (4-shot)26.3Math-Verify equivalence, Minerva prompts, 5,000 problems ove40.7

Interactive version: theaggregate.ai/model?slug=villanova2-2b-base · How It Works · Data refreshed daily, snapshot 2026-11-16.