mercor/ApexAgentsRecipe-TBench2_1-EvalTraces download history

mercor/ApexAgentsRecipe-TBench2_1-EvalTraces is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 688 times (429 in the last 7 days), and 794 times in total. It ranks #25,722 among datasets by monthly downloads.

ApexAgents Recipe — Terminal-Bench 2.1 Traces (k=3) Agent trajectories on Terminal-Bench 2.1 (89 tasks) for four model configurations: Qwen3.5-397B and Qwen3.6-35B, each before and after RL post-training. Each set contains 3 rollouts per task (267 trajectories per set). Failed and err

Open mercor/ApexAgentsRecipe-TBench2_1-EvalTraces on Hugging Face