Qwen-1.8B: benchmark results
Alibaba Cloud's Qwen-1.8B base model (Tongyi Qianwen), pretrained on over 2.2 trillion tokens with an 8,192-token context and a roughly 150,000-token vocabulary. Provider: Alibaba. Released 2023-11-30. Access: Open.
Aggregate Rating 1242 ± 34, rank #1546 of 1656 rated models, from 11 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ARC Challenge (AI2) | 53.2 | Accuracy (%) | 45.4 |
| CFLUE | 38.68 | Accuracy | 31.2 |
| C-Eval | 54.7 | Average (%) | 28.5 |
| BoolQ | 68 | Accuracy (%) | 23.4 |
| GSM8K | 21.2 | Accuracy (%) | 21.5 |
| InfiBench | 23.12 | Score (%) | 17.1 |
| PIQA | 73.3 | Accuracy (%) | 8.3 |
| MMLU | 28.2 | Accuracy (%) | 4.5 |
| Epoch AI - ECI | 93.07 | ECI Score | 4 |
| Epoch AI - Lambada | 58.4 | Score | 4 |
| Big-Bench Hard | 28.2 | Average (%) | 0 |
Interactive version: theaggregate.ai/model?slug=qwen-1-8b · How It Works · Data refreshed daily, snapshot 2026-10-20.