XIAOMI · MIMO 2.6
MiMo 2.6 Flash
For everyday conversations, content processing, and lightweight tool tasks. Use the unified Chat Completions format to integrate text conversations, reasoning, and function tools into existing applications.
- Model identifier
- mimo-v2.6-flash
- API protocol
- OpenAI Chat Completions
- Input and output
- Text messages and text results
- Interaction methods
- Standard responses, SSE, function tools
Choose the right model for your task
MiMo 2.6 Flash
For everyday conversations, content processing, and lightweight tool tasks. Evaluate application performance based on task quality, response time, and actual usage.
Another option in the same series
Pro is for complex tasks, while Flash is for everyday conversations and lightweight tasks. Different models are billed separately.
View MiMo 2.6 ProText conversations and reasoning
Pass conversation content through messages. Use enabled or disabled in thinking.type to control reasoning; reasoning content and the final answer are placed in reasoning_content and content, respectively.
max_completion_tokens is used to limit the total budget for reasoning and final output. Set an appropriate budget based on the task to avoid overly short limits truncating results.
Stream responses progressively
Set stream=true to receive SSE data. Applications can process reasoning deltas and answer deltas separately, then complete UI updates after receiving the end event.
To receive streaming usage data, set stream_options.include_usage=true and read the final usage information for logging.
Connect function tools in your application
Declare type=function tools and use tool_choice=auto. The application is responsible for executing the function and then returning the execution result to the model.
When continuing, preserve reasoning_content, complete tool_calls, and the corresponding tool_call_id in the same assistant message to keep the conversation history consistent.
Billing based on actual token usage
Input, output, and cached input tokens are calculated separately. Cached input is deducted from regular input usage; reasoning is already included in output tokens and is not billed again.
Package redemption ratios affect actual costs. Check current pricing and usage records to set a budget for your application.
View current billing rulesGetting Started
After activating the MiMo service and creating API credentials, send requests to https://api.acedata.cloud/mimo/chat/completions. The base_url of the OpenAI SDK can be set to https://api.acedata.cloud/mimo.
The unified /v1/chat/completions endpoint also requires the corresponding service permissions. See the integration guide for API parameters, examples, and error descriptions.
Frequently Asked Questions
Can I continue using the OpenAI SDK?
You can use the Chat Completions client by setting MiMo's base URL, API credentials, and the exact model identifier. Refer to this service's API documentation for specific parameters.
Is reasoning content charged repeatedly?
Reasoning tokens are already included in output usage and are charged once at the output price.
Are functions executed directly by the model?
Function tools are executed by your application. After returning the result along with the matching tool call identifier, continue generating the answer.
Are image input or built-in web search supported?
The current service scope covers text chat, reasoning, streaming responses, and function tools; it does not include multimodal input or built-in web search.
Can I use Responses, Messages, or Batch?
This service uses the Chat Completions API and does not provide these protocols or UltraSpeed models.
How can I check availability?
Check the console service status and call response information. When encountering rate limits or temporary unavailability, handle them according to the error description and avoid continuously resubmitting requests.
