Yi-1.5-34B-32K: benchmark results
01.AI's Yi-1.5-34B-32K, a 32K-context base model continuously pretrained from Yi on 500B additional high-quality tokens and fine-tuned on 3M diverse samples. Provider: 01.AI. Released 2024-05-15. Access: Open.
Aggregate Rating 1426 ± 21, rank #1123 of 1656 rated models, from 171 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - MMLU | 78.15 | Accuracy (%) (5-shot) | 99.2 |
| Open LLM Leaderboard v1 - WinoGrande | 83.98 | Accuracy (%) (5-shot) | 91.5 |
| Open LLM Leaderboard - GPQA | 36.33 | Score | 91.3 |
| Open Portuguese LLM - BLUEX | 67.87 | Accuracy (%) | 90.5 |
| Open Chinese LLM - CMMLU | 70.96 | Accuracy (%) | 89.9 |
| Open Chinese LLM - C-Eval Semantic | 88.81 | Accuracy (%) | 89.1 |
| Open Portuguese LLM - OAB Exams | 54.62 | Accuracy (%) | 86.8 |
| Open LLM Leaderboard - MMLU-Pro | 47.09 | Score | 86 |
| Open Chinese LLM - WinoGrande | 68.11 | Accuracy (%) | 85.8 |
| Open Portuguese LLM - ENEM | 73.55 | Accuracy (%) | 85.3 |
| Open Chinese LLM Leaderboard | 66.84 | Official Nonzero-Task Mean (%) | 84.9 |
| Open Chinese LLM - HellaSwag | 67.48 | Accuracy (%) | 83.4 |
Interactive version: theaggregate.ai/model?slug=yi-1-5-34b-32k · How It Works · Data refreshed daily, snapshot 2026-10-20.