List your company

Reasoning on BIG-Bench Hard (5 subtasks full)

86AccuracyThinking
13.82432.56251.370.038May 8, 2026

Evaluation results

MethodLinks
ThinkingBudget=2048, Backbone=Qwen3-8B, Thinking Mode=True2026.0586-11
NothinkBudget=1024, Backbone=Qwen3-8B, Thinking Mode=False2026.0575.11.4
NothinkBudget=2048, Backbone=Qwen3-8B, Thinking Mode=False2026.0575.1-11
ThinkingBudget=1024, Backbone=Qwen3-8B, Thinking Mode=True2026.0573.61.4
NothinkBudget=512, Backbone=Qwen3-8B, Thinking Mode=False2026.0566.420.6
NothinkBudget=256, Backbone=Qwen3-8B, Thinking Mode=False2026.0549.933.3
ThinkingBudget=512, Backbone=Qwen3-8B, Thinking Mode=True2026.0545.820.6
ThinkingBudget=256, Backbone=Qwen3-8B, Thinking Mode=True2026.0516.633.3