Qwen-14B-Chat: benchmark results
Alibaba's chat-aligned version of the original Qwen-14B (September 2023), pretrained on 3T multilingual tokens under the Tongyi Qianwen license. Provider: Alibaba. Released 2023-09-25. Access: Open.
Aggregate Rating 1416 ± 14, rank #1169 of 1656 rated models, from 146 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LLMsPark - Who Is Spy | 77 | Source-defined Who Is Spy game score (higher is better) | 100 |
| OpenCompass Language - Content Summarization - Chinese (CompassBench 2404) | 32 | Score (%) | 96.7 |
| OpenCompass Language - Translation (CompassBench 2404) | 19.2 | Score (%) | 90 |
| OpenCompass Language - Chinese Semantic Understanding - Chinese (CompassBench 2404) | 63.3 | Score (%) | 86.7 |
| OpenCompass Language - Intention Recognition - Chinese (CompassBench 2404) | 65.3 | Score (%) | 83.3 |
| NPHardEval - EDP | 26.91 | Accuracy (%) | 81.8 |
| NPHardEval - GCP D | 56 | Accuracy (%) | 81.8 |
| NPHardEval - SAS | 70.55 | Accuracy (%) | 81.8 |
| OpenCompass Knowledge - Natural Science - Chinese (CompassBench 2404) | 49.3 | Score (%) | 80 |
| OpenCompass Knowledge - Social Science - Chinese (CompassBench 2404) | 64.6 | Score (%) | 80 |
| OpenCompass LLM - Knowledge - Chinese (CompassBench 2404) | 56.4 | Score (%) | 80 |
| OpenCompass LLM - Language - Chinese (CompassBench 2404) | 55.5 | Score (%) | 80 |
Interactive version: theaggregate.ai/model?slug=qwen-14b-chat · How It Works · Data refreshed daily, snapshot 2026-10-20.