TNLG v2 (530B): benchmark results
Provider: Other.
Aggregate Rating 1316 ± 35, rank #8416 of 10916 rated models, from 32 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM Classic - CNN/DailyMail | 16.09 | ROUGE-2 (%) | 97.6 |
| HELM Classic - XSUM | 16.88 | ROUGE-2 (%) | 97.6 |
| HELM Classic - Dyck | 75.27 | Exact Match (%) | 95.6 |
| HELM Classic - NaturalQuestions Closed Book | 38.39 | F1 (%) | 87.9 |
| HELM | 78.65 | Mean win rate (imported) | 87.7 |
| HELM Classic - Entity Matching | 88.1 | Exact Match (%) | 86.4 |
| HELM Classic - MS MARCO Regular | 37.72 | RR@10 (%) | 86.2 |
| HELM Classic - MS MARCO TREC | 64.32 | NDCG@10 (%) | 83.3 |
| HELM Classic - BBQ | 47.93 | Exact Match (%) | 82.9 |
| HELM Classic - LSAT | 22.75 | Exact Match (%) | 82.4 |
| HELM Classic - Synthetic Reasoning Abstract | 36.25 | Exact Match (%) | 82.4 |
| HELM Classic - HellaSwag | 79.9 | Exact Match (%) | 80.6 |
Interactive version: theaggregate.ai/model?slug=tnlg-v2-530b · How It Works · Data refreshed daily, snapshot 2026-11-08.