List your company

Question Answering on MMLU-Pro (Natural Setting Test)

87.8AccuracySA (Base)
64.71270.70676.782.694Jan 29, 2026

Evaluation results

MethodLinks
SA (Base)Backbone=Gemini 3 Pro2026.0187.8
MA-Outcome (AG)Backbone=Gemini 3 Pro2026.0187.8
MA-Reasoning (ECL E)Backbone=Qwen 3-30B2026.0186.7
MA-Outcome (ECL I)Backbone=Deepseek V3.22026.0186.7
MA-Reasoning (AG)Backbone=GPT-5-mini2026.0186.7
MA-Reasoning (ECL E)Backbone=GPT-5-mini2026.0186.7
MA-Outcome (ECL E)Backbone=GPT-5.22026.0186.7
MA-Reasoning (ECL E)Backbone=GPT-5.22026.0186.7
MA-Outcome (ECL I)Backbone=Gemini 3 Flash2026.0186.7
MA-Reasoning (ECL E)Backbone=Gemini 3 Pro2026.0186.7
MA-Reasoning (ECL I)Backbone=Qwen 3-4B2026.0185.6
MA-Outcome (ECL E)Backbone=GPT-5-mini2026.0185.6
MA-Reasoning (ECL I)Backbone=GPT-5-mini2026.0185.6
MA-Outcome (AG)Backbone=GPT-5.22026.0185.6
MA-Reasoning (AG)Backbone=GPT-5.22026.0185.6
SA (Base)Backbone=Gemini 3 Flash2026.0185.6
MA-Outcome (ECL E)Backbone=Gemini 3 Flash2026.0185.6
MA-Reasoning (ECL E)Backbone=Gemini 3 Flash2026.0185.6
MA-Outcome (ECL I)Backbone=Gemini 3 Pro2026.0185.6
MA-Outcome (ECL E)Backbone=Gemini 3 Pro2026.0185.6
MA-Reasoning (AG)Backbone=Gemini 3 Pro2026.0185.6
MA-Reasoning (ECL I)Backbone=Gemini 3 Pro2026.0185.6
MA-Reasoning (ECL E)Backbone=Qwen 3-4B2026.0184.4
MA-Outcome (ECL E)Backbone=Deepseek V3.22026.0184.4
MA-Outcome (ECL I)Backbone=GPT-5-mini2026.0184.4
SA (Base)Backbone=GPT-5.22026.0184.4
MA-Reasoning (ECL I)Backbone=GPT-5.22026.0184.4
MA-Reasoning (ECL I)Backbone=Gemini 3 Flash2026.0184.4
MA-Reasoning (AG)Backbone=Deepseek V3.22026.0183.3
MA-Reasoning (ECL I)Backbone=Deepseek V3.22026.0183.3
MA-Reasoning (ECL E)Backbone=Deepseek V3.22026.0183.3
MA-Outcome (ECL I)Backbone=GPT-5.22026.0183.3
MA-Reasoning (AG)Backbone=Gemini 3 Flash2026.0183.3
MA-Reasoning (AG)Backbone=Qwen 3-30B2026.0182.2
MA-Reasoning (ECL I)Backbone=Qwen 3-30B2026.0182.2
MA-Outcome (AG)Backbone=GPT-5-mini2026.0182.2
MA-Outcome (AG)Backbone=Gemini 3 Flash2026.0182.2
MA-Outcome (AG)Backbone=Deepseek V3.22026.0181.1
SA (Base)Backbone=Deepseek V3.22026.0180
SA (Base)Backbone=GPT-5-mini2026.0180
MA-Reasoning (AG)Backbone=Qwen 3-4B2026.0178.9
MA-Outcome (ECL E)Backbone=Qwen 3-8B2026.0177.8
MA-Outcome (ECL I)Backbone=Qwen 3-30B2026.0177.8
MA-Outcome (ECL E)Backbone=Qwen 3-30B2026.0177.8
MA-Reasoning (ECL E)Backbone=Qwen 3-8B2026.0176.7
SA (Base)Backbone=Qwen 3-30B2026.0175.6
MA-Outcome (AG)Backbone=Qwen 3-30B2026.0173.3
MA-Outcome (ECL I)Backbone=Qwen 3-4B2026.0172.2
MA-Reasoning (ECL I)Backbone=Qwen 3-8B2026.0172.2
MA-Outcome (ECL E)Backbone=Qwen 3-4B2026.0171.1
SA (RL)Backbone=Qwen 3-8B2026.0171.1
MA-Reasoning (AG)Backbone=Qwen 3-8B2026.0171.1
SA (RL)Backbone=Qwen 3-4B2026.0170
MA-Outcome (AG)Backbone=Qwen 3-4B2026.0167.8
MA-Outcome (AG)Backbone=Qwen 3-8B2026.0166.7
MA-Outcome (ECL I)Backbone=Qwen 3-8B2026.0165.6