ResearchBenchmarksTroubleshooting on TS FloDialFollow7.5Worst Case Interaction LengthGoT7.2928.69610.111.504Feb 2, 2026Evaluation ResultsMethodMethodLinksWorst Case Interaction LengthGoTBackbone=Qwen 2.5 72B...Backbone=Qwen 2.5 72B Instruct2026.027.5GoTBackbone=GPT 4.1Backbone=GPT 4.12026.027.9UoTBackbone=GPT 4.1Backbone=GPT 4.12026.029UoTBackbone=Qwen 2.5 72B...Backbone=Qwen 2.5 72B Instruct2026.029DCBackbone=Qwen 2.5 72B...Backbone=Qwen 2.5 72B Instruct2026.0210.5DPBackbone=Qwen 2.5 72B...Backbone=Qwen 2.5 72B Instruct2026.0210.7DCBackbone=GPT 4.1Backbone=GPT 4.12026.0211.6DPBackbone=GPT 4.1Backbone=GPT 4.12026.0212.7