granite-3.0-3B-a800m-base: benchmark results
IBM's Granite 3.0 3B-A800M Base, a mixture-of-experts model (3B total, about 800M active) trained in two stages: 8 trillion tokens from diverse domains, then 2 trillion tokens curated. Provider: IBM. Released 2024-10-21. Access: Open.
Aggregate Rating 1307 ± 30, rank #1464 of 1656 rated models, from 48 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Dutch NLU - DBRD | 89.16 | Matthews correlation coefficient (%) | 63.8 |
| EuroEval Portuguese NLU - SST-2 PT | 77.24 | Matthews correlation coefficient (%) | 51.4 |
| EuroEval Spanish NLU - MLQA ES | 59.26 | Character F1 (%) | 48.2 |
| EuroEval Italian NLU - SQuAD IT | 63.37 | Character F1 (%) | 44.3 |
| EuroEval Portuguese NLU | 45.29 | NLU Average Score (%) | 39.4 |
| EuroEval Spanish NLU - ScaLA ES | 4.28 | Matthews correlation coefficient (%) | 37.6 |
| EuroEval Portuguese NLU - MultiWikiQA PT | 62.9 | Character F1 (%) | 35.9 |
| EuroEval Portuguese NLU - HAREM | 36.07 | Micro F1 excluding MISC (%) | 34.8 |
| EuroEval Portuguese NLU - ScaLA PT | 4.95 | Matthews correlation coefficient (%) | 34.3 |
| EuroEval Portuguese | 32.3 | Average Score (%) | 30.6 |
| BenCzechMark - NER | 27.27 | Duel Win Score (%) | 28.6 |
| EuroEval Spanish NLU | 32.84 | NLU Average Score (%) | 28.3 |
Interactive version: theaggregate.ai/model?slug=granite-3-0-3b-a800m-base · How It Works · Data refreshed daily, snapshot 2026-10-20.