comma-v0.1-1t: benchmark results
Provider: Common Pile. Released 2025-05-15. Access: Open.
Aggregate Rating 1284 ± 32, rank #8923 of 10916 rated models, from 104 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval English NLU - SQuAD | 82.38 | Character F1 (%) | 68.1 |
| EuroEval Norwegian Knowledge - Idioms NO | 2.5 | MCC (x100) | 44.9 |
| EuroEval Faroese NLU - ScaLA FO | 2 | Matthews correlation coefficient (%) | 43.8 |
| EuroEval Italian NLU - Sentipolc16 | 44.34 | Matthews correlation coefficient (%) | 38.8 |
| EuroEval Portuguese NLU - MultiWikiQA PT | 64.99 | Character F1 (%) | 38.5 |
| EuroEval Spanish NLU - MLQA ES | 54.75 | Character F1 (%) | 37.7 |
| EuroEval Spanish NLU - ScaLA ES | 4.18 | Matthews correlation coefficient (%) | 37.2 |
| EuroEval Icelandic NLU - ScaLA IS | 1.06 | Matthews correlation coefficient (%) | 36.8 |
| EuroEval Italian NLU - SQuAD IT | 54.82 | Character F1 (%) | 35.6 |
| EuroEval French NLU - FQuAD | 58.1 | Character F1 (%) | 34.9 |
| EuroEval Spanish NLU - Sentiment Headlines ES | 31.75 | Matthews correlation coefficient (%) | 34 |
| EuroEval Danish NLU - Angry Tweets | 36 | Matthews correlation coefficient (%) | 32.6 |
Interactive version: theaggregate.ai/model?slug=comma-v0-1-1t · How It Works · Data refreshed daily, snapshot 2026-11-08.