List your company

Agentic Task Performance on AssistantBench (test)

65.8Easy AccuracyLearning to Share
42.71248.70654.760.694Feb 5, 2026

Evaluation results

MethodLinks
Learning to ShareShared Memory=true, Model=Qwen3 (32B)2026.0265.825.214.420.31,479
Learning to ShareShared Memory=true, Model=GPT-5.12026.026135.12026.7882
M1-ParallelShared Memory=false, Model=GPT-5.12026.0257.335.216241,389
MagenticOneShared Memory=false, Model=Qwen3 (32B)2026.0253.117.48.513.41,084
M1-ParallelShared Memory=false, Model=Qwen3 (32B)2026.0252.621.48.514.72,239
MagenticOneShared Memory=false, Model=GPT-5.12026.0243.631.515.321.7724