Terminus-4B：一个较小的模型能否在智能体执行任务中取代前沿大语言模型？
Terminus-4B: Can a Smaller Model Replace Frontier LLMs at Agentic Execution Tasks?
AI总结：
研究智能体终端执行任务中较小模型能否取代前沿大语言模型。提出经监督微调及强化学习训练的Terminus-4B模型，评估发现其能减少主智能体令牌使用量约30%，不影响性能，还缩小与前沿模型差距甚至超越其性能。
摘要：
现代编码智能体越来越多地将专门子任务委托给子智能体，子智能体是处理诸如搜索、调试或终端执行等狭窄职责的小型、专注的智能体循环。这种架构模式通过在子智能体上下文中隔离冗长输出（如构建日志、测试结果等）来保持主智能体的上下文窗口干净。通常智能体在执行此类任务时使用前沿模型作为子智能体。本文研究了经过微调的小语言模型（SLM）在智能体终端执行任务中能否达到与前沿模型相当的性能。我们展示了Terminus-4B，它是通过基于规则的大语言模型评判奖励的监督微调（SFT）和强化学习（RL）对Qwen3-4B模型进行训练后的模型，专门用于此任务。在广泛评估中，我们发现Terminus-4B与无子智能体基线相比，能将主智能体的令牌使用量减少约30%，且不影响在SWE-Bench Pro和内部SWE-Bench C#基准测试等基准上的智能体性能。此外，Terminus-4B改善了关键指标，表明主智能体更多依赖子智能体输出，自身执行的终端任务更少。我们的模型不仅缩小了普通Qwen模型与Claude Sonnet / Opus / GPT-5.3-Codex等前沿模型之间的差距，甚至常常超越它们的性能。
英文摘要：
Modern coding agents increasingly delegate specialized subtasks to subagents, which are smaller, focused agentic loops that handle narrow responsibilities like search, debugging or terminal execution. This architectural pattern keeps the main agent's context window clean by isolating verbose outputs (e.g. build logs, test results, etc.) within the subagent context. Typically when agents employ subagents for such tasks, they use frontier models as these subagents. In this paper, we investigate whether a finetuned small language model (SLM) can achieve comparable performance to frontier models in the task of agentic terminal execution. We present Terminus-4B, which is a post-trained Qwen3-4B model via Supervised Finetuning (SFT) and Reinforcement Learning (RL) using rubric-based LLM-as-judge reward, specifically for this task. In our extensive evaluation spanning various frontier models, training ablations and main agent configurations, we find that Terminus-4B is able to reduce the token usage of the main agent by up to ~30% compared to the No Subagent baseline with no impact to agent performance on benchmarks like SWE-Bench Pro and our internal SWE-Bench C# benchmark, which tends to be heavy in verbose execution tasks. Furthermore, Terminus-4B improves key metrics showing the main agent relying on the outputs of the subagent and doing fewer terminal execution tasks by itself. We see that our model not only closes the gap between the Vanilla Qwen model and frontier models like Claude Sonnet / Opus / GPT-5.3-Codex, but often even exceeds their performance.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
