qwen-code vs mini-swe-agent
AI coding agents compared
qwen-code and mini-swe-agent use language models to help with software engineering tasks. qwen-code offers a broad set of interfaces, model providers, and agent features, while mini-swe-agent centers on a compact Python agent that acts through bash.

qwen-code: Build with Qwen Models Across Coding Interfaces
Qwen Code is an agentic coding tool built around Qwen models, designed to work from the terminal and through editor, desktop, browser, and chat interfaces. It also supports other model providers and local models, with SDKs and automation modes for custom workflows.

mini-swe-agent: Automate Software Engineering Tasks with an AI Agent
mini-swe-agent is a compact Python agent that uses a language model and bash to tackle software engineering tasks, including GitHub issues. Its simple control flow suits developers who want a hackable CLI or a baseline for agent research.
| qwen-code | mini-swe-agent | |
|---|---|---|
| Language | TypeScript | Python |
| License | Apache-2.0 | MIT |
| Stars | 28.6k | 8.2k |
| Forks | 3.2k | 1.1k |
| Last analyzed | Oct 19, 2026 | Oct 3, 2026 |
Key differences
- qwen-code is written in TypeScript and provides TypeScript, Python, and Java SDKs; mini-swe-agent is written in Python and offers Python bindings.
- qwen-code supports terminal, editor, desktop, browser, and chat interfaces; mini-swe-agent provides a CLI and Python bindings.
- qwen-code includes built-in workflows and tools such as plan mode, hooks, skills, LSP integration, and MCP; mini-swe-agent uses bash as its agent interface.
- qwen-code supports Qwen, OpenAI, Anthropic, Gemini, third-party providers, and local models; mini-swe-agent lists integrations including LiteLLM, OpenRouter, and Portkey.
- qwen-code is licensed under Apache-2.0; mini-swe-agent is licensed under MIT.
- qwen-code offers headless execution and an HTTP/SSE daemon mode, with the web UI and daemon described as experimental; mini-swe-agent supports local and several sandboxed execution environments.
Choose qwen-code if you…
- need coding-agent workflows across terminal, editors, desktop, browser, or chat.
- want to switch among supported model providers or use local models within one workflow.
- plan to use TypeScript, Python, or Java SDKs, or run tasks in scripts and CI/CD.
Choose mini-swe-agent if you…
- want a compact, hackable Python agent that uses bash for actions.
- need a simple scaffold for comparing or researching language models and inspecting agent trajectories.
- plan to run tasks in local or sandboxed environments using the CLI or Python bindings.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.