Qwen1.5-MoE-A2.7B-Chat: benchmark results
Qwen1.5-MoE-A2.7B-Chat is a mixture-of-experts model with 2.7B active parameters per token, as reported by its publisher. Provider: Alibaba. Released 2024-03-28. Access: Open.
Aggregate Rating 1335 ± 16, rank #1417 of 1656 rated models, from 113 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Chinese LLM - CMMLU | 54.31 | Accuracy (%) | 55.9 |
| RewardBench Chat Hard | 63.16 | Accuracy (%) | 55.1 |
| Open Chinese LLM - HellaSwag | 60.57 | Accuracy (%) | 54.1 |
| Open Portuguese LLM - ASSIN2 RTE | 89.74 | Macro F1 (%) | 51.9 |
| Open Portuguese LLM - BLUEX | 52.02 | Accuracy (%) | 51.4 |
| Open Portuguese LLM - OAB Exams | 42.32 | Accuracy (%) | 51.4 |
| Open LLM Leaderboard v1 - MMLU | 60.97 | Accuracy (%) (5-shot) | 50.4 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 50.56 | MC2 (%) (0-shot) | 48.7 |
| Open Portuguese LLM - FaQuAD NLI | 63.93 | Macro F1 (%) | 48.2 |
| Open Chinese LLM - WinoGrande | 62.59 | Accuracy (%) | 45.7 |
| Open LLM Leaderboard v1 - GSM8K | 28.2 | Accuracy (%) (5-shot) | 45.7 |
| Open Chinese LLM - C-Eval Semantic | 67.78 | Accuracy (%) | 44.4 |
Interactive version: theaggregate.ai/model?slug=qwen1-5-moe-a2-7b-chat · How It Works · Data refreshed daily, snapshot 2026-10-20.