MiMo on RunAPI

Use OpenAI-compatible Chat Completions at https://runapi.ai/v1 as the primary protocol. Keep the exact public model identity throughout the request.

Primary protocol recipe

Authenticate

Set OPENAI_API_KEY to a RunAPI API key and OPENAI_BASE_URL to https://runapi.ai/v1. Keep the key in the environment or a secret manager.

Send request

from openai import OpenAI

client = OpenAI(api_key="YOUR_RUNAPI_TOKEN", base_url="https://runapi.ai/v1")
response = client.chat.completions.create(
    model="mimo-v2.5-pro",
    messages=[{"role": "user", "content": "Summarize this decision."}],
)
print(response.choices[0].message.content)

For longer output, request terminal usage while streaming:

stream = client.chat.completions.create(
    model="mimo-v2.5",
    messages=[{"role": "user", "content": "Write a status update."}],
    stream=True,
    stream_options={"include_usage": True},
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
    if chunk.usage:
        print(chunk.usage)

Verify result

For a synchronous call, require one response choice with a final assistant message and read usage from that response. For SSE, consume through [DONE] and require the terminal usage chunk requested by include_usage. An HTTP 2xx without the expected final content and Usage is incomplete.

Stop boundaries

Correct a rejected request shape at most once, using the structured error and this verified allowlist. Retry a transport failure once only when no response or Usage was returned and replay is safe. On a terminal RunAPI failure, retain the request ID and error, then stop without changing model or protocol.

Positive verified request allowlist

Start with only this verified request shape:

• mimo-v2.5-pro: a text-only messages history with system, user, and assistant roles.
• mimo-v2.5: the same text shape, plus synchronous Chat Completions content parts containing text and public HTTP(S) image_url.url values.
• Both models: synchronous calls or SSE with stream_options.include_usage.
• mimo-v2.5 Chat Completions: an optional metadata object with string values, for example {"user_id": "customer-123"}, on synchronous or streaming requests. Do not rely on metadata being echoed in the response or available for stored-completion lookup.
• mimo-v2.5 Chat Completions supports custom function tools, tool_choice (auto, none, required, or a named function), and parallel_tool_calls. Preserve assistant tool_calls, their IDs, and reasoning_content when supplied; return each result as a tool message with the matching tool_call_id. Continue until the model returns final text. Both sync and SSE are supported.
• For Claude Code with mimo-v2.5, use native Anthropic Messages as described in the compatibility reference. Tool calling is not enabled for Pro or Responses.

Add an optional control only when the current RunAPI contract or a successful validation result explicitly verifies it for the selected model and mode.

Compatibility protocols

Load compatibility protocols only when an existing client requires Responses or Anthropic Messages. Keep the primary recipe in this file for new integrations.

Supported models

References

• Model overview and pricing: https://runapi.ai/models/mimo.md
• Provider page: https://runapi.ai/providers/xiaomi.md
• Catalog: https://runapi.ai/models.md