Xiaomi MiMo
Xiaomi MiMo's open-source terminal coding agent with cross-session memory, subagents, goal-driven loops and OpenCode compatibility.
Quick answer
MiMoCode is a distinct coding-agent product rather than another MiMo model card: it packages memory, orchestration, tools and provider routing into an installable open-source CLI.
MiMoCode is a terminal-native AI coding agent released by Xiaomi MiMo. It can read and write code, run commands and manage Git, while storing project memory, checkpoints, notes and task progress across sessions with SQLite FTS5 search. Built as an OpenCode fork, it retains multi-provider, TUI, LSP, MCP and plugin support and adds context reconstruction, parallel subagents, a goal-completion judge, spec-driven Compose workflows, voice input and Dream/Distill commands. Stable v0.1.13 adds scoped permissions, checkpoint and Skill disable controls, stable MCP tool ordering, an OpenAI-compatible capability API, ALL_TOOLS support, Codex mode, child-environment isolation and removal of an unnecessary GPT context cap.
Editorial verdict
Developers who want an open terminal coding agent focused on durable project memory, long-running tasks and configurable model providers.
Avoid adopting it without reviewing the additional use restrictions, command permissions, memory files and hosted-provider data terms.
MiMoCode is a distinct coding-agent product rather than another MiMo model card: it packages memory, orchestration, tools and provider routing into an installable open-source CLI.
Open-source CLI; MiMo Auto is free for a limited time, while hosted and custom-provider costs vary
Free open-source installation, MiMo Auto limited-time access, Xiaomi MiMo Platform OAuth, Custom provider API billing
Commercial use should follow the current product, API, model license and billing terms.
Review prompt, file, media upload, retention and training-use terms before sensitive workloads.
Use project memory, checkpoints and task progress to resume large codebase work without rebuilding all context.
Compose mode bundles planning, TDD, debugging, review, verification and merge-oriented workflows.
Subagents can work in parallel while the goal judge checks whether the stopping condition has actually been met.
Choose MiMo, import Claude Code authentication or configure another OpenAI-compatible provider in the TUI.
Checked against stable v0.1.13 on August 28, 2026. Main-branch changes newer than that release are not treated as shipped; installation, free access, provider support and restrictions can change quickly.
qwen-code
opencode-go
kimi-code
codebuddy
Official · EN · verified 2026-08-04
Documents installation, agents, memory, context management, subagents, goal judging, Compose, voice, providers and OpenCode ancestry.
Official · EN · verified 2026-06-11
Official MiMo product route linked from the repository.
Terms · EN · verified 2026-06-11
Lists prohibited military, malicious cyber, privacy-violating and unsupervised high-risk uses.
Official · EN · verified 2026-08-28
Confirms the current stable version, scoped permissions, checkpoint and Skill switches, stable MCP ordering, capability API, Codex mode and child-environment isolation.
Last checked: 2026-08-28
No approved reviews yet.