List your company
ResearchFormerly WizWand

Multi-task Language Understanding on MMLU-Pro (Accuracy, AVG., Improvement Overhead)

72.47AccuracyQwen3-235B-A22B
44.046851.425958.80566.1841Aug 19, 2025
Updated 7mo ago

Evaluation results

MethodLinks
Qwen3-235B-A22BFramework=Reference Model, Backbone Model=Qwen3-235B-A22B2025.0872.4778.22—
COCO Qwen3-8B with coco(Llama-3.1-8B)Framework=COCO Framework (Proposed), Backbone Model=Qwen3-8B, Monitor Model=Llama-3.1-8B2025.0868.6974.376.5
COCO Qwen3-8B with coco(Qwen3-8B)Framework=COCO Framework (Proposed), Backbone Model=Qwen3-8B, Monitor Model=Qwen3-8B2025.0866.674.186.2
Aflow-Qwen3-8BFramework=Multi-Agent Framework Baseline, Backbone Model=Qwen3-8B2025.0866.5669.86—
Qwen3-8BFramework=Reference Model, Backbone Model=Qwen3-8B2025.0858.8568.52—
COCO Llama-3.1-8B with coco(Qwen3-8B)Framework=COCO Framework (Proposed), Backbone Model=Llama-3.1-8B, Monitor Model=Qwen3-8B2025.0853.4463.599.5
Llama-3.1-8BFramework=Reference Model, Backbone Model=Llama-3.1-8B2025.0848.0355.48—
COCO Llama-3.1-8B with coco(Llama-3.1-8B)Framework=COCO Framework (Proposed), Backbone Model=Llama-3.1-8B, Monitor Model=Llama-3.1-8B2025.0845.6258.460.63
Aflow-Llama3.1-8BFramework=Multi-Agent Framework Baseline, Backbone Model=Llama-3.1-8B2025.0845.1458.09—