AIJun 2

What Makes Interaction Trajectories Effective for Training Terminal Agents?

arXiv:2606.0346151.6h-index: 24
AI Analysis

For researchers in agentic AI post-training, this work challenges the assumption that stronger teachers are always better, introducing a data-efficient paradigm based on harness engineering.

The paper investigates whether stronger code agents are better teachers for post-training, finding that lower-scoring agents can produce better student generalization due to Environment-Grounded Supervision (EGS). With only 15.3k trajectories, Qwen3-32B achieves 24.3% on Terminal-Bench 2.0, rivaling prior SOTA using over 30x more data.

Stronger code agents are commonly assumed to be superior teachers for post-training, yet this assumption remains poorly disentangled from task difficulty, harness design, and student capacity. We investigate this pedagogical link using Terminal-Lego, a scalable pipeline that transforms multi-domain real-world issues into environment-verified agentic tasks. Surprisingly, standalone performance does not dictate teaching efficacy: while Claude Opus 4.6 achieves higher scores on Terminal-Bench 2.0, students fine-tuned on trajectories from DeepSeek-V3.2, a lower-scoring agent, exhibit significantly stronger generalization. We attribute this "pedagogical paradox" to Environment-Grounded Supervision (EGS): trajectories that explicitly expose inspect-act-verify behaviors through harness-visible interactions allow students to internalize robust problem-solving routines rather than fragile action sequences. Scaling analysis reveals exceptional data efficiency: with only 15.3k Terminal-Lego trajectories, for example, Qwen3-32B achieves a 24.3% score on Terminal-Bench 2.0, rivaling previous SOTA performance established with over 30x the data volume. Our results suggest that the frontier of agent post-training lies beyond mere outcome-matching, shifting the focus toward "Harness Engineering", where the systematic design of environment-grounded interaction structures serves as the primary catalyst for reproducible and generalizable agentic intelligence.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes