Xiaomi MiMo Provider for Claude Code
1T MoE model at a fraction of frontier pricing
claude-multi add mimoUse cases
MiMo-V2.5-Pro is a 1 trillion parameter mixture-of-experts model with 42B active parameters and a 1M token context window. It delivers frontier-class coding performance at a significantly lower cost per token than comparable models. The Anthropic-compatible endpoint means it drops straight into Claude Code.
Model specs
| Role | Model | Parameters | Context | Max Output |
|---|---|---|---|---|
| Primary (Opus/Sonnet) | MiMo-V2.5-Pro | 1T MoE (42B active) | 1M | 128K |
| Fast (Haiku) | MiMo-V2.5 | 310B MoE (15B active) | 1M | 128K |
V2.5-Pro handles the heavy reasoning. V2.5 runs background tasks, subagent calls, and quick edits at lower cost. The template maps them to the right roles automatically.
Thinking mode is enabled by default.
Setup
- Create an account at xiaomimimo.com and generate an API key
- Run the setup command:
claude-multi add mimo- Paste your API key when prompted
The template sets up the base URL, model mappings, and thinking parameters. Your instance is ready to use immediately.
Token Plan alternative
Xiaomi also offers a subscription-based Token Plan with a monthly credit pool. If you prefer predictable monthly costs, use the mimo-token template instead. Note that Token Plan uses different regional endpoints (CN, SG, EU). After setup, edit ~/.claude-<name>/settings.json and replace ANTHROPIC_BASE_URL with the endpoint from your subscription console.
When to pick MiMo
MiMo is the cost-performance leader. If you do a lot of coding and want frontier quality without frontier pricing, MiMo-V2.5-Pro delivers. The 1M context window handles large codebases, and the MoE architecture means you only pay for the active parameters.
For developers who prefer a fixed monthly bill, switch to the MiMo Token Plan template.
Pricing details
MiMo is available two ways: pay-per-token at xiaomimimo.com, or a subscription Token Plan (monthly credit pool, regional CN/SG/EU endpoints — use the mimo-token template). The 1T MoE architecture keeps per-token cost low because only 42B of the 1T parameters are active per request. Check xiaomimimo.com for current rates.
Related providers
Pay-per-token via xiaomimimo.com