Trae Agent is an LLM-based agent for general purpose software engineering tasks.

Harbor is a framework for running agent evaluations and creating and using RL environments.

Convert GitHub PRs into Harbor tasks

[ICLR2026🔥Oral] SwingArena: Competitive Programming Arena for Long-context GitHub Issue Solving

Edit Banana: A framework for converting statistical formats into editable.

Advances and Frontiers of LLM-based Issue Resolution in Software Engineering A Comprehensive Survey

The Github repo for our survey paper: "Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models"

DART-GUI: Efficient Multi-turn RL for GUI Agents via Decoupled Training and Adaptive Data Curation

SWE-Lego: Pushing the Limits of Supervised Fine-tuning for Software Issue Resolving

💻 Terminal-Agent with Human-in-the-Loop Learning

[ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training

Coding problems used in aider's polyglot benchmark

A unified suite for generating elite reasoning problems and training high-performance LLMs, including pioneering attention-free architectures

Repo2Run is an LLM-based agent that automates environment configuration by generating error-free Dockerfiles for Python repositories.

[ICLR 2026] MMSearch-Plus: Benchmarking Provenance-Aware Search For Multimodal Browsing Agents

Tongyi Deep Research, the Leading Open-source Deep Research Agent

SWE-Flow: Synthesizing Software Engineering Data in a Test-Driven Manner

The 100 line AI agent that solves GitHub issues or helps you in your command line. Radically simple, no huge configs, no giant monorepo—but scores >74% on SWE-bench verified!

An AI Hedge Fund Team

Democratizing Reinforcement Learning for LLMs

🙌 OpenHands: AI-Driven Development

[NeurIPS 2025 D&B Spotlight] Scaling Data for SWE-agents

Reproducing R1 for Code with Reliable Rewards

SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024]