Llama-Poro-2-70B-SFT: benchmark results
Provider: Meta. Released 2025-05-01. Access: Open.
Aggregate Rating 1620 ± 33, rank #835 of 10916 rated models, from 114 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval German NLU - Sb10K | 62.15 | Matthews correlation coefficient (%) | 98.5 |
| EuroEval Finnish NLU - Scandisent FI | 93.25 | Matthews correlation coefficient (%) | 98.1 |
| EuroEval Swedish NLU - Swerec | 80.46 | Matthews correlation coefficient (%) | 97.9 |
| EuroEval Finnish NLU | 68.35 | NLU Average Score (%) | 96.8 |
| EuroEval Faroese NLU - FoQA | 77.78 | Character F1 (%) | 96.4 |
| EuroEval Danish NLU - Angry Tweets | 59.1 | Matthews correlation coefficient (%) | 96.3 |
| EuroEval Norwegian NLU - Norec | 62.46 | Matthews correlation coefficient (%) | 95.5 |
| EuroEval Finnish NLU - ScaLA FI | 45.6 | Matthews correlation coefficient (%) | 95.4 |
| EuroEval English | 76.26 | Average Score (%) | 94.9 |
| EuroEval Italian NLU - Sentipolc16 | 66.56 | Matthews correlation coefficient (%) | 92.6 |
| EuroEval Italian | 64.31 | Average Score (%) | 92.3 |
| EuroEval Italian NLU - MultiNERD IT | 81.08 | Micro F1 excluding MISC (%) | 92.3 |
Interactive version: theaggregate.ai/model?slug=llama-poro-2-70b-sft · How It Works · Data refreshed daily, snapshot 2026-11-08.