Yi-34B-200K: benchmark results
01.AI's 34B bilingual base model trained on 3T tokens (November 2023), with its context window extended to 200K. Provider: 01.AI. Released 2023-11-06. Access: Open.
Unified ELO 1423 ± 1, rank #1058 of 1392 rated models, from 43 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - GPQA | 14.21 | Score | 89.8 |
| Open Chinese LLM - WinoGrande | 69.14 | Accuracy (%) | 89.3 |
| Open LLM Leaderboard - MMLU-Pro | 39.27 | Score | 84.3 |
| Open Chinese LLM - CMMLU | 68.37 | Accuracy (%) | 80.7 |
| Open Chinese LLM - HellaSwag | 66.01 | Accuracy (%) | 80.4 |
| Open Chinese LLM - C-Eval Semantic | 85.72 | Accuracy (%) | 77.7 |
| Open Chinese LLM - ARC Challenge | 57.17 | Accuracy (%) | 74.2 |
| Open LLM Leaderboard - BBH | 36.02 | Score | 73.5 |
| HELMET (128K) | 40.3 | Average Score | 71 |
| EffiBench - NET | 2.91 | Normalized Execution Time | 67.1 |
| MERA - ruModAr | 70.15 | EM (%) | 63.2 |
| BABILong (NIAH) | 54 | Avg Accuracy (%) | 62.1 |
Interactive version: theaggregate.ai/model?slug=yi-34b-200k · How It Works · Data refreshed daily, snapshot 2026-09-05.