# Open-SWE-Traces

> NVIDIA dataset of software-engineering agent trajectories (OpenHands, SWE-agent, mini-swe-agent, Claude Code, OpenCode) for distilling SWE agents; CC BY 4.0.

- id: `open-swe-traces`
- category: Information (`information`)
- url: https://huggingface.co/datasets/nvidia/Open-SWE-Traces
- status: active
- pricing: free
- license: CC-BY-4.0
- tags: dataset, agent-trajectories, software-engineering, training-data, huggingface
- auth: none
- access notes: Not gated; load with `datasets.load_dataset("nvidia/Open-SWE-Traces", "v2.0")` or download the parquet files.
- related: SWE-rebench (https://indexagentica.com/entries/swe-rebench/); SWE-smith (https://indexagentica.com/entries/swe-smith/); SWE-Gym (https://indexagentica.com/entries/swe-gym/); OpenHands (https://indexagentica.com/entries/openhands/); mini-swe-agent (https://indexagentica.com/entries/mini-swe-agent/)
- sources: https://huggingface.co/datasets/nvidia/Open-SWE-Traces
- added: 2026-10-27
- updated: 2026-10-27
- last verified: 2026-10-27
- featured in: Agent environment frameworks compared: Harbor vs AgentEnv Framework vs Seahaven (https://indexagentica.com/compare/agent-environment-frameworks/)
- maintainer: NVIDIA
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/open-swe-traces/
- json: https://indexagentica.com/api/entries/open-swe-traces.json

Full message histories, tool definitions, model and gold patches and success metadata for issue-resolution tasks drawn from permissively licensed task sets (Scale-SWE, SWE-rebench-V2). Versioned configs: v1.0 (151k cleaned trajectories from MiniMax and Qwen models; git-hacking runs removed), v1.1 (DeepSeek-V4-Flash, Qwen3.6-27B), v1.2 (Qwen3.8-27B) and v2.0, released 2026-10-26 (Qwen3.8-Flash-Next, xhigh reasoning) with splits per harness. Size category 100K–1M rows; parquet. The card states it is ready for commercial and non-commercial use.