Language Understanding on MMLU (50 samples per category stratified sampling)
79.18AccuracyGoA_Max
Evaluation results
| Method | Links | |
|---|---|---|
| GoA_MaxAgent Configuration=Multi-Agent, Number of Agents=3, Zero-shot CoT=true2026.04 | 79.18 | |
| GoA_MeanAgent Configuration=Multi-Agent, Number of Agents=3, Zero-shot CoT=true2026.04 | 78.52 | |
| Self-MoAAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 78.14 | |
| SCAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 77.97 | |
| Qwen2.5-7B-InstructAgent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=General2026.04 | 77.61 | |
| RefineAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 77.4 | |
| MoAAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 75.71 | |
| DebateAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 72.53 | |
| ReConcileAgent Configuration=Multi-Agent, Number of Agents=6, Zero-shot CoT=true2026.04 | 69.61 | |
| Qwen2.5-Coder-7B-InstructAgent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=Code2026.04 | 68.04 | |
| Mathstral-7B-v0.1Agent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=Math2026.04 | 63.47 | |
| Saul-7B-Instruct v1Agent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=Legal2026.04 | 55.86 | |
| finance-Llama3-8BAgent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=Finance2026.04 | 54.11 | |
| Bio-Medical-Llama-3-8BAgent Configuration=Single-Agent, Number of Agents=1, Zero-shot CoT=true, Agent Role/Strategy=Biomedical2026.04 | 46.6 |