DeepSWE

DeepSWE is a benchmark for measuring frontier coding agents on original, long-horizon software engineering tasks drawn from active open-source repositories. The benchmark includes 113 tasks across TypeScript, Go, Python, JavaScript, and Rust, with isolated environments and program-based verifiers.

DeepSWE tasks use the Harbor task format:

task.toml         Metadata: repository, base commit, language, prebuilt image, resource limits
instruction.md    The prompt the agent sees
environment/      Dockerfile that reproduces the prebuilt image (fallback if the image is unavailable)
tests/            Verifier: test.sh (entry point) + test.patch (test additions, applied at grading time)
solution/         Reference solution (held out from the agent; for human and AI reviewers)

The verifier exercises the behavior the prompt describes. It accepts any solution whose observable behavior is correct, regardless of internal symbol names or structure. The reference patch in solution/ is never used at grading time; it exists so reviewers can spot-check correctness offline.

Use Pier to run the benchmark:

git clone https://github.com/datacurve-ai/deep-swe
uv tool install datacurve-pier

# Claude Opus 4.7 via Claude Code
export ANTHROPIC_API_KEY=...
pier run -p deep-swe/tasks --agent mini-swe-agent --model anthropic/claude-opus-4-7

# GPT-5.5 via Codex
export OPENAI_API_KEY=...
pier run -p deep-swe/tasks --agent mini-swe-agent --model openai/gpt-5.5

Pier is a Harbor-compatible framework for sandboxed coding-agent evals. It began as a fork of Harbor to support CLI agents in air-gapped tasks: Harbor blocks all outbound traffic in allow_internet = false tasks, including dependency installs and LLM API calls. Pier adds per-agent network allowlists, giving agents only the network access they need while keeping the task environment isolated.

Pier also adds more complete trajectory metadata, a better trajectory viewer, and pier critique run for analyzing agent trajectories. All leaderboard scores were produced with Pier running mini-swe-agent on Modal.

mini-swe-agent is model-agnostic. Pier also drives claude-code, codex, gemini-cli, and opencode directly. Pass --env modal to run in parallel sandboxes on Modal.

Deterministic random subset of the 113-task corpus:

pier run -p deep-swe/tasks --agent mini-swe-agent --n-tasks 10 --sample-seed 0

Single task:

pier run -p deep-swe/tasks/<task-id> --agent mini-swe-agent

This repository includes a Bun/TypeScript browser for reviewing the task corpus locally:

bun run task-browser

Open http://localhost:4173 to search tasks, read the agent prompt, inspect task metadata, and review the reference solution or verifier patch in a GitHub-style diff viewer. Set PORT=... to use a different localhost port.