mercor/ApexAgentsRecipe-TBench2_1-EvalTraces

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 errored rollouts are included and count as failures. Companion dataset: ApexAgentsRecipe-ApexAgents480-EvalTraces (the same four models evaluated on the 480-task APEX-Agents benchmark).… See the full description on the dataset page: https://huggingface.co/datasets/mercor/ApexAgentsRecipe-TBench2_1-EvalTraces.

<iframe src="https://silytics.si/embed/item/ec0aa78a-48ca-4f72-ad51-3f605a1b1628" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.