opt-13B: benchmark results
Meta's OPT-13B, part of the Open Pretrained Transformer suite of decoder-only models from 125M to 175B parameters, designed to match GPT-3-class model sizes and performance. Provider: Meta. Released 2022-05-03. Access: Open.
Aggregate Rating 1172 ± 31, rank #1607 of 1656 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenBookQA | 39.8 | Accuracy (%) | 26.2 |
| HellaSwag | 69.9 | Accuracy (%) | 25 |
| Open LLM Leaderboard v1 - WinoGrande | 68.51 | Accuracy (%) (5-shot) | 22.6 |
| Open LLM Leaderboard v1 - HellaSwag | 71.2 | Normalized accuracy (%) (10-shot) | 21.3 |
| WinoGrande | 64.7 | Accuracy (%) | 21.3 |
| BoolQ | 65 | Accuracy (%) | 18.2 |
| Open LLM Leaderboard v1 - GSM8K | 1.74 | Accuracy (%) (5-shot) | 18.2 |
| Open Portuguese LLM - FaQuAD NLI | 43.97 | Macro F1 (%) | 17.1 |
| PIQA | 75.7 | Accuracy (%) | 16.7 |
| Open LLM Leaderboard v1 - ARC Challenge | 39.93 | Normalized accuracy (%) (25-shot) | 16.2 |
| Open Portuguese LLM - ASSIN2 RTE | 46.65 | Macro F1 (%) | 14.9 |
| Open Portuguese LLM - OAB Exams | 24.51 | Accuracy (%) | 13.2 |
Interactive version: theaggregate.ai/model?slug=opt-13b · How It Works · Data refreshed daily, snapshot 2026-10-20.