ResearchBenchmarksGeneral Reasoning on General Reasoning AggregateFollow45.4Accuracy (Macro)TAC17.00824.37931.7539.121Jun 23, 2026Evaluation ResultsMethodMethodLinksAccuracy (Macro)TACBackbone=Qwen3-4B-Base...Backbone=Qwen3-4B-Base, Seeds=12026.0645.4SECBackbone=Qwen3-4B-Base...Backbone=Qwen3-4B-Base, Seeds=12026.0643.8RandomBackbone=Qwen3-4B-Base...Backbone=Qwen3-4B-Base, Seeds=12026.0643.2TACBackbone=Qwen3-0.6B-Ba...Backbone=Qwen3-0.6B-Base, Seeds=12026.0621.6SECBackbone=Qwen3-0.6B-Ba...Backbone=Qwen3-0.6B-Base, Seeds=12026.0619.9RandomBackbone=Qwen3-0.6B-Ba...Backbone=Qwen3-0.6B-Base, Seeds=12026.0618.1