Skip to content
agentboards
Compare/goose vs mini-SWE-agent

goosevsmini-SWE-agent

Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.

goose
Agentic AI Foundation (originally Block) · Terminal agent
#14OSSMCP
Panel
6.8
5 spec wins
Reliability
6.3
Usefulness
6.5
Cost
6.8
Longevity
7.3

“Block gave it to the Linux Foundation, which is the one exit strategy that cannot be un-acquired.”

mini-SWE-agent
SWE-agent (Princeton and Stanford) · Terminal agent
#8OSS
Panel
6.9
0 spec wins
Reliability
6.3
Usefulness
5.5
Cost
8.8
Longevity
7.0

“One hundred lines of Python, which is fewer than most competitors spend on their pricing page.”

Spec by spec

Specgoosemini-SWE-agent
Architecture
CategoryTerminal agentTerminal agent
Runslocallocal, sandbox
Platformsmacos, linux, windowsmacos, linux
Context windownot documentednot documented
Protocols
MCP clientYesNo
MCP serverYesNo
Capabilities
Runs terminal commandsYesYes
Multi-file editsYesYes
Git operationsYesYes
Browser controlYesThe built-in Computer Controller extension drives macOS apps including the browser through the Peekaboo CLI, rather than via a dedicated CDP or Playwright tool. No
Sandboxed executionYes`goose session|run --container <id>` executes extensions inside an existing Docker container; goose itself runs unsandboxed on the host otherwise. Yes
Multi-agent orchestrationYesNo
Headless / CI modeYesYes
Models
BackboneClaude, GPT, Gemini, Ollama (local), OpenRouter, Azure OpenAI, Amazon Bedrockany via litellm
Bring your own modelYesYes
Local modelsYesYes
Cost
Pricing modelbyokbyok
Starts at$0/mon/a
Free tierYesYes
Bring your own keyYesYes
Openness
Open sourceYesYes
LicenseApache-2.0MIT
GitHub stars55,1748,410

Which one would each critic pick

Criticgoosemini-SWE-agentPick
El Juez——not enough reviews
El Amigo7.36.8goose — A free, open, local agent under the Linux Foundation that talks to any model and any MCP server; you will spend a weekend tuning it and then own it.
El Crítico6.56.5no preference
El Profesor6.58.0mini-SWE-agent — The reference harness for the SWE-bench bash-only leaderboard: one bash action per turn via subprocess.run, a linear history, no tool-calling API, and Gemini 3 Pro reported above 74% on Verified.
La Inversora5.36.8mini-SWE-agent — No company, a Princeton and Stanford lab with 6,938 stars; the funding is grants and the exit is a paper, which is more stable than half the cap tables on this board.
La Jefa5.85.0goose — Free, local, Apache-2.0, with a foundation instead of a vendor; the security questionnaire has nobody to send it to, and sixty engineers get sixty keys.
El Hacker9.38.5goose — Rust, Apache-2.0, every extension an MCP server in YAML, Ollama on my box, a foundation instead of a vendor; this is the one I would build if I had the time.

Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.

Want a third column? The compare tool handles any two agents. Three-way comparisons are on the roadmap once the spec rows are all verified.
↑↓ navigate · Enter open · Esc close589 agents + guides