AIarXiv
Ranked by formulaVerdict: RoutineTev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards
Long-horizon agents have drawn growing attention for their ability to work autonomously -- from hours to days -- on economically valuable tasks, a capability essential to automated discovery and recursive…
Caveats
- Preprint; not yet peer reviewed.
VerdictCompetent work. Briefs at most.
Abstract
Long-horizon agents have drawn growing attention for their ability to work autonomously -- from hours to days -- on economically valuable tasks, a capability essential to automated discovery and recursive self-improvement (RSI). Yet open-source models still struggle on long-horizon tasks, largely for lack of both training data and a working recipe, and scaling both is hard: long, complex tasks are difficult to curate and verify, and their rewards are sparse. We address both. First, we introduce Tev-Gym, the first long-horizon agent training dataset with verifiable, dense, progressive rewards -- 1,170 verified task environments across 7 domains. It is built on a latent-world-first approach that first constructs a latent world: a complete and correct hidden specification of the task domain, workflow, and expected state after each step. Each task is then progressively rendered from this specification into an agent-facing environment, enabling controllable complexity and a natural progress metric. Second, we turn this metric into a training recipe: via progress-potential reward shaping (PPRS), we convert progress into a dense per-step reward that improves training efficiency and makes otherwise-intractable tasks trainable. Training with this recipe yields Tev-27B, which achieves substantial gains on the Agents' Last Exam computing and math sciences domains (ALE-CM), matching GPT-6 Astra and Claude Opus 5 at Pass@5. Despite training only on ALE-related data, it generalizes to the Terminal-Bench (TB) 3.0 and 4.0 subsets, surpassing GLM 5.2 and Gemini 3.7 Flash, and reaches Claude Opus 4.8-level performance on middle-horizon tasks such as TB 2.1.
The editor's rubric
| Dimension | Level | Weight | What that level means |
|---|---|---|---|
| Leverage | ████░ 4 | 24% | A general-purpose tool used across several fields (Adam, ResNet, LoRA, next-generation sequencing). |
| Magnitude | ███░░ 3 | 18% | Large gain: roughly 2x, or a clear new state of the art on a hard, unsaturated problem. |
| Evidence | ██░░░ 2 | 14% | Limited: single setting, weak baselines, or an observational association presented as causal. |
| Novelty | ███░░ 3 | 16% | A genuinely new approach to an open problem. |
| Trajectory | ███░░ 3 | 18% | A clear path to scale. |
| Stakes | ██░░░ 2 | 10% | Benefits a professional community (practitioners, clinicians, engineers). |
Editor’s rationale
Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: method (we propose); breadth (generalizes, many tasks); gains (outperforms); novelty (discovery); verification (code released); scale (scalable, efficient); stakes (general AI).
How the score was computed
- Merit
- 4.9 / 10
- Adjusted merit
- 4.4 / 10
- Attention
- 88%
- Freshness
- 69%
- Hugging Face upvotes95 (reference 25, via hf-daily, Oct 15, 2026, 02:05 UTC)