arXiv:2609.05576v1 Announce Type: new Abstract: The paradigm of LLMs has rapidly shifted from passive language interfaces to autonomous Claw-like agents that execute long-horizon tasks across statefu...
The paper introduces Pufibara, an agent harness designed to maintain engineering state and evidence across revisions in Modelica-based physical system modeling. It also presents a 232-task Modelica Agent Workflow Benchmark covering model repair, generation, and tuning, evaluated by an external benchmark-owned evaluator. Experiments show Pufibara outperforms Claude Code in task success and resource efficiency across two LLM backends.
arXiv:2610.08720v1 Announce Type: new Abstract: LLM-based agents are increasingly advancing scientific and engineering problem solving, with physics simulation emerging as a challenging yet practical...
arXiv:2604. 06742v2 Announce Type: replace-cross Abstract: The evolution of Large Language Models (LLMs) has catalyzed a paradigm shift towards intent-driven software development, where autonomous agents are expected to design and deliver complete, runnable software systems from scratch.
The paper introduces TerminalHorizon, an automated engine that synthesizes long‑horizon trajectories for terminal agents by reconstructing environments from real‑world professional work and progressively expanding tasks to require further exploration and improvement. Using this engine, the authors create 1.5K task environments and fine‑tune Qwen3.5‑27B, producing TerminalHorizon‑27B, which shows significant performance gains across 12 benchmarks in coding, agentic production tasks, and reasoning. Case studies demonstrate the model’s ability to experiment beyond initial solutions, refine methods based on feedback, and preserve effective improvements in final deliverables.
ProgramGym introduces a scalable framework for training long-horizon program reproduction agents, comprising Program-Env and Program-Agent components. Program-Env builds verifiable training environments from repositories via a feedback loop that repairs sandbox builds and improves black-box tests, yielding 1,648 environments across C, C++, Go, and Rust. Program-Agent employs a role-isolated workflow of analysis, implementation, and review with revision management, boosting performance on ProgramBench and raising Qwen3.8-27B’s accuracy from 34.87% to 42.84% and surpassing flagship models with fine-tuning.
arXiv:2608. 05466v1 Announce Type: new Abstract: High-quality long-horizon training data for terminal agents is expensive to produce, often costing hundreds to thousands of dollars per task, because each task must keep the instruction, environment, reference solution, and verifier mutually consistent.
arXiv:2606. 14672v1 Announce Type: new Abstract: Large language models increasingly serve as execution engines for agentic systems, yet they still consume context through a sequential text interface.
E2E-SWE is a benchmark that tests large language models’ ability to create complete, functional software repositories from scratch. It includes 186 tasks across 11 programming languages, each requiring an agent to build an installable project based solely on a natural‑language specification and an empty workspace, while passing a hidden test suite. The benchmark was crafted by software engineers and LLMs, then refined through iterative verification by autonomous agents to ensure clarity and solvability.
SPADE (Self-Play in Adaptive Synthetic Executable Environments) is a reinforcement‑learning framework where a single large language model acts as both an Environment Designer—creating executable, long‑horizon training environments—and a Reasoning Agent—learning to act within those environments. The framework uses a regret signal based on the difference between rewarded performance with and without privileged hints to guide the Designer toward environments that are challenging yet solvable. Experiments show that, when scaled to 30‑billion‑parameter models, SPADE outperforms fixed‑environment baselines by significant margins across math, science, code, and reasoning benchmarks, and improves tool‑use performance on BFCL‑v4 and ACEBench‑Agent. whyItMatters":"By making environment design a learnable component, SPADE enables continuous self‑improvement and demonstrates that adaptive, self‑generated training environments can substantially boost language‑model performance across diverse tasks."
The paper proposes adding a verification layer to large language model agents, using a V-model design where each step of the agent’s plan has a dedicated verifier. In a pilot study on the MuSiQue dataset, verified configurations outperformed unverified ones, with deterministic gates correcting most errors at zero cost. The authors plan to extend the architecture to full integration and evaluate it over the next year.
The paper introduces a benchmark for evaluating large language model (LLM) agents in closed‑loop embedded software development. It provides five embedded‑control tasks with realistic build‑and‑runtime feedback, testing agents that must translate requirements into code, self‑verify, and iterate until device behavior meets specifications. Seven GPT‑family and Qwen‑family models were evaluated across four feedback scenarios, revealing that gpt‑5.4 leads in pass rate but does not fully saturate the benchmark, while qwen3.5‑27B is the strongest local model and smaller local models perform poorly.