Baichuan-2-13B-Base: benchmark results
Provider: Baichuan. Released 2023-09-06. Access: Open.
Unified ELO 1353 ± 1, rank #1264 of 1392 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| GSM8K | 52.8 | Accuracy (%) | 53.8 |
| Big-Bench Hard | 49 | Average (%) | 47.9 |
| Epoch AI - Lambada | 74 | Score | 44 |
| MMLU | 59.2 | Accuracy (%) | 32.6 |
| PIQA | 78.1 | Accuracy (%) | 31.7 |
| HellaSwag | 70.8 | Accuracy (%) | 30.9 |
| BoolQ | 67 | Accuracy (%) | 22.1 |
| InfiBench | 26.32 | Score (%) | 21.9 |
| ARC Challenge (AI2) | 38 | Accuracy (%) | 15.8 |
Interactive version: theaggregate.ai/model?slug=baichuan-2-13b-base · How It Works · Data refreshed daily, snapshot 2026-09-05.