Curie (6.7B): benchmark results
Provider: Other.
Aggregate Rating 1085 ± 31, rank #10852 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Classic - BLiMP | 83.15 | Exact Match (%) | 75.8 |
| HELM Classic - Entity Data Imputation | 81.06 | Exact Match (%) | 68.2 |
| HELM Classic - BBQ | 36.23 | Exact Match (%) | 58.5 |
| HELM Classic - Synthetic Reasoning Abstract | 22.33 | Exact Match (%) | 51.5 |
| HELM Classic - TruthfulQA | 23.19 | Exact Match (%) | 49.2 |
| HELM Classic - CivilComments | 53.86 | Exact Match (%) | 45.5 |
| HELM Classic - WikiFact | 23.58 | Exact Match (%) | 43.9 |
| HELM Classic - Synthetic Reasoning Natural | 14.89 | F1 (%) | 42.6 |
| RealToxicityPrompts | 0.06 | Toxic fraction | 40.2 |
| HELM Classic - Dyck | 47 | Exact Match (%) | 30.9 |
| HELM Classic - QuAC | 32.11 | F1 (%) | 30.8 |
| HELM Classic - XSUM | 9.08 | ROUGE-2 (%) | 29.3 |
Interactive version: theaggregate.ai/model?slug=curie-6-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.