ResearchBenchmarksColor Querying on CoP-QA-FFollow1AC ScoreTalk2DM0.8976640.9242320.95080.977368Feb 12, 2026Evaluation ResultsMethodMethodLinksAC ScoreAQ ScoreTalk2DMLLM Backbone=Qwen3:30BLLM Backbone=Qwen3:30B2026.0211Talk2DMLLM Backbone=GPT-oss:20BLLM Backbone=GPT-oss:20B2026.0210.9979Talk2DMLLM Backbone=Magistral...LLM Backbone=Magistral:24B2026.0210.8629Talk2DMLLM Backbone=Gemma3:27BLLM Backbone=Gemma3:27B2026.0211Talk2DMLLM Backbone=Llama3.1:8BLLM Backbone=Llama3.1:8B2026.020.99790.8755Talk2DMLLM Backbone=Deepseek-...LLM Backbone=Deepseek-r1:32B2026.020.90160.8962