granite-swash-3B-a600m: benchmark results

Provider: IBM. Access: Open.

Aggregate Rating 1306 ± 33, rank #8548 of 10552 rated models, from 26 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Lara Mini Base Report (base-model protocol) - PIQA (0-shot)77.4Accuracy normalized by answer length (%), English benchmark51.9
Lara Mini Base Report (base-model protocol) - XNLI (English) (0-shot)53.2Accuracy (%), English benchmark51.9
Lara Mini Base Report (base-model protocol) - WinoGrande (5-shot)66.1Accuracy (%), English benchmark43.5
Lara Mini Base Report (base-model protocol) - HellaSwag (10-shot)69.2Accuracy normalized by answer length (%), English benchmark42.6
Lara Mini Base Report (base-model protocol) - MATH (4-shot)25.8Math-Verify equivalence, Minerva prompts, 5,000 problems ove38.9
Lara Mini Base Report (base-model protocol) - Global PIQA (parallel), All Languages (0-shot)23.9Accuracy (%) averaged over all languages/variants of the ben38
Lara Mini Base Report (base-model protocol) - XNLI, All Languages (0-shot)38.9Accuracy (%) averaged over all languages/variants of the ben37
Lara Mini Base Report (base-model protocol) - COPA (0-shot)80.2Accuracy (%), English benchmark35.2
Lara Mini Base Report (base-model protocol) - MMLU (5-shot)54.4Accuracy (%), English benchmark35.2
Lara Mini Base Report (base-model protocol) - Global MMLU, All Languages (5-shot)37.4Accuracy (%) averaged over all languages/variants of the ben31.5
Lara Mini Base Report (base-model protocol) - INCLUDE, All Languages (5-shot)35.6Accuracy (%) averaged over all languages/variants of the ben31.5
Lara Mini Base Report (base-model protocol) - ARC-Challenge (25-shot)49.6Accuracy normalized by answer length (%), English benchmark27.8

Interactive version: theaggregate.ai/model?slug=granite-swash-3b-a600m · How It Works · Data refreshed daily, snapshot 2026-11-16.