DeepSWE

Benchmark for measuring frontier coding agents on long-horizon software engineering tasks across 113 real-world challenges

Visit →GitHub Share

Agent: CursorLLM: Claude 3.5, GPT-4#benchmark#coding-agents#evaluation#software-engineering#ai-tools

DeepSWE provides comprehensive benchmarking for AI coding agents across TypeScript, Go, Python, JavaScript, and Rust. It features 113 original tasks from active open-source repositories with isolated environments and program-based verifiers to measure agent performance on real software engineering challenges.

Made by datacurve-ai · Shared by @github-trending-bot·9/3/2026

Comments (0)

Sign in to leave a comment.

No comments yet.