Apertus-v1.5-8B-base: benchmark results

Provider: Apertus. Access: Open.

Aggregate Rating 1438 ± 33, rank #4427 of 10552 rated models, from 26 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Lara Mini Base Report (base-model protocol) - Global PIQA (parallel), All Languages (0-shot)30.1Accuracy (%) averaged over all languages/variants of the ben100
Lara Mini Base Report (base-model protocol) - Global PIQA (non-parallel), All Languages (0-shot)64.3Accuracy (%) averaged over all languages/variants of the ben98.1
Lara Mini Base Report (base-model protocol) - T-HellaSwag, All Languages (0-shot)52.7Accuracy normalized by answer length (%) averaged over all l98.1
Lara Mini Base Report (base-model protocol) - XCOPA, All Languages (0-shot)65.2Accuracy (%) averaged over all languages/variants of the ben98.1
Lara Mini Base Report (base-model protocol) - XNLI, All Languages (0-shot)44.6Accuracy (%) averaged over all languages/variants of the ben94.4
Lara Mini Base Report (base-model protocol) - ARC-Challenge (25-shot)64.5Accuracy normalized by answer length (%), English benchmark92.6
Lara Mini Base Report (base-model protocol) - INCLUDE, All Languages (5-shot)59.3Accuracy (%) averaged over all languages/variants of the ben92.6
Lara Mini Base Report (base-model protocol) - Global MMLU, All Languages (5-shot)56.3Accuracy (%) averaged over all languages/variants of the ben90.7
Lara Mini Base Report (base-model protocol) - MGSM (8 (CoT)-shot)54.6Exact match on each language's native CoT answer phrase, ave90.7
Lara Mini Base Report (base-model protocol) - COPA (0-shot)87.8Accuracy (%), English benchmark89.8
Lara Mini Base Report (base-model protocol) - HellaSwag (10-shot)80.3Accuracy normalized by answer length (%), English benchmark89.8
Lara Mini Base Report (base-model protocol) - PIQA (0-shot)80.1Accuracy normalized by answer length (%), English benchmark88.9

Interactive version: theaggregate.ai/model?slug=apertus-v1-5-8b-base · How It Works · Data refreshed daily, snapshot 2026-11-16.