StackMap
Subscribe

mini-swe-agent vs openinterpreter

The 100-line agent from the SWE-bench team: >74% on SWE-bench Verified with no tools but bash, no config sprawl — the reference minimal harness, adopted by Meta, NVIDIA and Ramp. — versus — Terminal coding agent optimized for low-cost open models — a Rust Codex fork with switchable harness emulation (claude-code, kimi-code, qwen-code…), ACP support and Codex SDK compatibility.

The curated verdict

Standalone open coding agents at two extremes: mini-swe-agent is the 100-line research baseline; Open Interpreter is a production CLI tuned per-model for cheap open weights.

mini-swe-agentopeninterpreter
Stars6.3k68k
Forks8665.8k
LanguagePythonRust
LicenseMITApache-2.0
Last activity6 days ago8 days ago
Topicscoding, agentscoding, agents
Curated connections63

mini-swe-agent — the curator's take

The existence proof that most harness complexity is optional: the team that built SWE-bench and SWE-agent asked what a 100x simpler agent loses — the answer is almost nothing (>74% Verified), which is why it became the standard baseline harness for benchmarking models (Ramp's SWE-bench, DeepSWE — where it beats Claude Code and Codex as a harness). Read it to understand agents; use it to evaluate models fairly. NOT a daily driver: no MCP, no skills, no IDE plumbing — by design. If you're choosing a tool to ship features with, this is the control group, not the product.

openinterpreter — the curator's take

The harness-emulation bet is unique: instead of one house style, it reimplements the provider-recommended harness per model (Kimi, Qwen, DeepSeek…), which is exactly what cheap open models need to perform. Codex-SDK-compatible, so switching costs one line. When NOT: if you live on Claude or GPT frontier models, their native CLIs stay ahead of the emulation; and the project has reinvented itself before (the original Python open-interpreter era) — expect direction changes.