OLMo 3 7B: benchmark results
Ai2's fully open OLMo 3 7B base model, trained on ~5.9T Dolma 3 tokens with ~65K context and the entire 'model flow' released (November 2025). Provider: Allen AI. Released 2025-11-20. Access: Open.
Aggregate Rating 1428 ± 13, rank #1112 of 1656 rated models, from 394 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval French NLU - Eltec | 76.8 | Micro F1 excluding MISC (%) | 98.9 |
| EuroEval English NLU - CoNLL EN | 81.75 | Micro F1 excluding MISC (%) | 88.6 |
| EuroEval Polish NLU - KPWr NER | 64.76 | Micro F1 excluding MISC (%) | 87.6 |
| EuroEval Ukrainian NLU - NER UK | 68.71 | Micro F1 excluding MISC (%) | 86.2 |
| EuroEval Hungarian NLU - Szeged NER | 71.37 | Micro F1 excluding MISC (%) | 84.2 |
| EuroEval Spanish NLU - CoNLL ES | 73.11 | Micro F1 excluding MISC (%) | 83.9 |
| EuroEval Portuguese NLU - HAREM | 51.69 | Micro F1 excluding MISC (%) | 80.9 |
| EuroEval Portuguese NLU - ScaLA PT | 25.37 | Matthews correlation coefficient (%) | 80 |
| EuroEval Norwegian NLU - NorNE NB | 75.14 | Micro F1 excluding MISC (%) | 78.2 |
| EuroEval Dutch NLU - CoNLL NL | 65.41 | Micro F1 excluding MISC (%) | 77.6 |
| EuroEval Bosnian | 50.8 | Average Score (%) | 75.2 |
| EuroEval Catalan NLU - WikiANN CA | 68.1 | Micro F1 excluding MISC (%) | 73.5 |
Interactive version: theaggregate.ai/model?slug=olmo-3-7b · How It Works · Data refreshed daily, snapshot 2026-10-20.