moe-7B-1B-active-shared-experts: benchmark results

Provider: ibm-research. Access: Open.

Aggregate Rating 1336 ± 34, rank #7991 of 10552 rated models, from 26 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Lara Mini Base Report (base-model protocol) - XNLI (English) (0-shot)55.8Accuracy (%), English benchmark92.6
Lara Mini Base Report (base-model protocol) - COPA (0-shot)87.6Accuracy (%), English benchmark86.1
Lara Mini Base Report (base-model protocol) - HellaSwag (10-shot)78.5Accuracy normalized by answer length (%), English benchmark85.2
Lara Mini Base Report (base-model protocol) - WinoGrande (5-shot)72.4Accuracy (%), English benchmark72.2
Lara Mini Base Report (base-model protocol) - MBPP+ (3-shot)53.2pass@1; must pass MBPP's own and the EvalPlus tests (%)68.5
Lara Mini Base Report (base-model protocol) - PIQA (0-shot)78.7Accuracy normalized by answer length (%), English benchmark67.6
Lara Mini Base Report (base-model protocol) - ARC-Challenge (25-shot)56.5Accuracy normalized by answer length (%), English benchmark61.1
Lara Mini Base Report (base-model protocol) - MBPP (3-shot)45.2pass@1 (execution) (%)58.3
Lara Mini Base Report (base-model protocol) - CRUXEval-O (2-shot)32.5pass@1, direct prediction, official answer parser (%)50.9
Lara Mini Base Report (base-model protocol) - PAWS (English) (0-shot)65Accuracy (%), English benchmark50
Lara Mini Base Report (base-model protocol) - PAWS-X, All Languages (0-shot)54.6Accuracy (%) averaged over all languages/variants of the ben47.2
Lara Mini Base Report (base-model protocol) - GSM8K (5-shot)54.9Exact match (strict extraction: first "#### N", else first \46.3

Interactive version: theaggregate.ai/model?slug=moe-7b-1b-active-shared-experts · How It Works · Data refreshed daily, snapshot 2026-11-16.