Yet Even Less Is Even Better For Agentic, Reasoning, and Coding LLMs

• URL: http://arxiv.org/abs/2604.00824v3
• Date: Mon, 06 Apr 2026 05:39:35 GMT
• Title: Yet Even Less Is Even Better For Agentic, Reasoning, and Coding LLMs
• Abstract summary: Training effective software engineering agents requires large volumes of task-specific trajectories.<n>We propose an end-to-end training framework that achieves superior agentic capabilities with fewer but higher-quality training trajectories.
• Score: 28.672417618778976
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Training effective software engineering agents requires large volumes of task-specific trajectories, incurring substantial data construction costs. Inspired by the "Less-Is-More" hypothesis in mathematical reasoning, we investigate its extension to agentic scenarios and propose an end-to-end training framework that achieves superior agentic capabilities with fewer but higher-quality training trajectories. This is achieved via STITCH (Sliding-memory Trajectory Inference and Task Chunking Heuristic), a coarse-to-fine mechanism that filters low-value noise and retains decision-critical tokens to maximize training signal quality. We conduct experiments across multiple agent frameworks (e.g., mini-SWE-agent, MSWE-agent), model scales (30B to 355B), and multilingual settings (Python, Java, and ArkTS). On SWE-bench Verified, models trained with STITCH achieve up to 63.16% relative improvement over base models. On Multi-SWE-bench (Java), MiniMax-M2.5-STITCH achieves 43.75% with our CodeArts Agent scaffold (+16.67%). On HarmonyOS (ArkTS), GLM-4.7-STITCH improves the compilation pass rate to 61.31% (+43.34%) with less than 1K training trajectories. Our results confirm that the "Less-Is-More" paradigm generalizes effectively to complex agentic tasks across diverse languages and model scales.

Related papers

• Orchard: An Open-Source Agentic Modeling Framework [124.68499958175111]
  We present Orchard, an open-source framework for scalable agentic modeling.<n>At its core is Orchard Env, a lightweight environment service providing reusable primitives for sandbox lifecycle management.<n>On top of Orchard Env, we build three agentic modeling recipes.
  arXiv  Detail & Related papers  (2026-05-14T16:35:12Z)
• AgentCPM-Explore: Realizing Long-Horizon Deep Exploration for Edge-Scale Agents [75.67445299298949]
  AgentCPM-Explore is a compact 4B agent model with high knowledge density and strong exploration capability.<n>We introduce a holistic training framework featuring parameter-space model fusion, reward signal denoising, and contextual information refinement.<n>AgentCPM-Explore achieves state-of-the-art (SOTA) performance among 4B-class models, matches or surpasses 8B-class SOTA models on four benchmarks, and even outperforms larger-scale models such as Claude-4.5-Sonnet or DeepSeek-v3.2 in five benchmarks.
  arXiv  Detail & Related papers  (2026-02-06T08:24:59Z)
• SWE-RM: Execution-free Feedback For Software Engineering Agents [61.86380395896069]
  Execution-based feedback is widely used in the development of coding agents through test-time scaling (TTS) and reinforcement learning (RL)<n>In contrast, execution-free feedback from reward models can provide more fine-grained signals without depending on unit test cases.<n>We introduce SWE-RM, an accurate and robust reward model adopting a mixture-of-experts architecture with 30B total parameters and 3B activated during inference.
  arXiv  Detail & Related papers  (2025-12-26T08:26:18Z)
• AgentMath: Empowering Mathematical Reasoning for Large Language Models via Tool-Augmented Agent [80.83250816918861]
  Large Reasoning Models (LRMs) like o3 and DeepSeek-R1 have achieved remarkable progress in natural language reasoning with long chain-of-thought.<n>However, they remain computationally inefficient and struggle with accuracy when solving problems requiring complex mathematical operations.<n>We present AgentMath, an agent framework that seamlessly integrates language models' reasoning capabilities with code interpreters' computational precision.
  arXiv  Detail & Related papers  (2025-12-23T19:57:49Z)
• Towards Engineering Multi-Agent LLMs: A Protocol-Driven Approach [13.760107452858044]
  This paper introduces Software Engineering Multi-Agent Protocol (SEMAP), a protocol-layer methodology that instantiates three core SE design principles for multi-agents.<n>In code development, it achieves up to a 69.6% reduction in total failures function-level development and 56.7% for deployment-level development.
  arXiv  Detail & Related papers  (2025-10-14T03:49:30Z)
• Kimi-Dev: Agentless Training as Skill Prior for SWE-Agents [43.20918899874477]
  Large Language Models (LLMs) are increasingly applied to software engineering (SWE)<n>We first curate the Agentless training recipe and present Kimi-Dev, an open-source SWE LLM achieving 60.4% on SWE-bench Verified.<n>With additional SFT adaptation on 5k publicly-available trajectories, Kimi-Dev powers SWE-Agents to 48.6% pass@1, on par with that of Claude 3.5 Sonnet.
  arXiv  Detail & Related papers  (2025-09-27T01:49:13Z)
• Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models [49.911784762244814]
  TraceRL is a trajectory-aware reinforcement learning framework for diffusion language models (DLMs)<n>We derive a series of state-of-the-art diffusion language models, namely TraDo.<n>TraDo-8B-Instruct achieves relative accuracy improvements of 6.1% over Qwen2.5-7B-Instruct and 51.3% over Llama3.1-8B-Instruct on mathematical reasoning benchmarks.
  arXiv  Detail & Related papers  (2025-09-08T17:58:06Z)
• Cooperative Multi-Agent Planning with Adaptive Skill Synthesis [16.228784877899976]
  We present a novel multi-agent architecture that integrates vision-language models (VLMs) with a dynamic skill library and structured communication for decentralized closed-loop decision-making.<n>The skill library, bootstrapped from demonstrations, evolves via planner-guided tasks to enable adaptive strategies.<n>We demonstrate its strong performance against state-of-the-art MARL baselines across both symmetric and asymmetric scenarios.
  arXiv  Detail & Related papers  (2025-02-14T13:23:18Z)
• Building Math Agents with Multi-Turn Iterative Preference Learning [56.71330214021884]
  This paper studies the complementary direct preference learning approach to further improve model performance.<n>Existing direct preference learning algorithms are originally designed for the single-turn chat task.<n>We introduce a multi-turn direct preference learning framework, tailored for this context.
  arXiv  Detail & Related papers  (2024-09-04T02:41:04Z)
• DS-Agent: Automated Data Science by Empowering Large Language Models with Case-Based Reasoning [56.887047551101574]
  We present DS-Agent, a novel framework that harnesses large language models (LLMs) agent and case-based reasoning (CBR) In the development stage, DS-Agent follows the CBR framework to structure an automatic iteration pipeline, which can flexibly capitalize on the expert knowledge from Kaggle. In the deployment stage, DS-Agent implements a low-resource deployment stage with a simplified CBR paradigm, significantly reducing the demand on foundational capabilities of LLMs.
  arXiv  Detail & Related papers  (2024-02-27T12:26:07Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.