List your company

Question Answering on GPQA Natural Setting (test)

90AccuracyMA-Reasoning (ECL I)
40.653.42566.2579.075Jan 29, 2026

Evaluation results

MethodLinks
MA-Reasoning (ECL I)Backbone=Qwen 3-30B2026.0190
MA-Reasoning (ECL E)Backbone=DeepSeek V3.22026.0187.5
MA-Reasoning (ECL E)Backbone=Qwen 3-30B2026.0185
MA-Outcome (ECL E)Backbone=DeepSeek V3.22026.0185
MA-Reasoning (AG)Backbone=DeepSeek V3.22026.0185
MA-Reasoning (AG)Backbone=Qwen 3-4B2026.0182.5
MA-Reasoning (ECL I)Backbone=Qwen 3-4B2026.0182.5
MA-Reasoning (AG)Backbone=Qwen 3-30B2026.0182.5
MA-Reasoning (ECL I)Backbone=DeepSeek V3.22026.0182.5
MA-Reasoning (ECL E)Backbone=Qwen 3-4B2026.0177.5
MA-Reasoning (AG)Backbone=Qwen 3-8B2026.0177.5
MA-Reasoning (ECL E)Backbone=Qwen 3-8B2026.0177.5
SA (Base)Backbone=DeepSeek V3.22026.0177.5
MA-Outcome (AG)Backbone=DeepSeek V3.22026.0177.5
MA-Reasoning (ECL I)Backbone=Qwen 3-8B2026.0175
MA-Outcome (ECL E)Backbone=Qwen 3-30B2026.0175
MA-Outcome (ECL I)Backbone=DeepSeek V3.22026.0175
SA (Base)Backbone=Qwen 3-30B2026.0172.5
MA-Outcome (ECL E)Backbone=Qwen 3-4B2026.0170
MA-Outcome (ECL I)Backbone=Qwen 3-30B2026.0170
MA-Outcome (ECL E)Backbone=Qwen 3-8B2026.0165
MA-Outcome (AG)Backbone=Qwen 3-30B2026.0165
MA-Outcome (ECL I)Backbone=Qwen 3-4B2026.0162.5
MA-Outcome (AG)Backbone=Qwen 3-8B2026.0157.5
SA (RL)Backbone=Qwen 3-4B2026.0155
MA-Outcome (ECL I)Backbone=Qwen 3-8B2026.0155
SA (RL)Backbone=Qwen 3-8B2026.0152.5
MA-Outcome (AG)Backbone=Qwen 3-4B2026.0142.5