Model API · Available
MiMo-V2.6-Flash, Xiaomi's mixture-of-experts model with 310B parameters and 12B active, through an OpenAI-compatible endpoint on hardware AxForge owns in Stockholm. Fast answers and tool calls for agents; it thinks when you ask it to.
Why AxForge
| OpenAI-compatible | Chat completions, streaming, tool calls and usage reporting follow the OpenAI API shape. Point your SDK at api.axforge.ai/v1 and pick the model mimo-v2.6-flash. |
|---|---|
| Built for agents | Only 12B of its 310B parameters work on each token, so answers come fast, and tool calls come back in the OpenAI shape, streamed or not. |
| Zero prompt retention | Prompts and completions are processed in memory in Stockholm, Sweden. Not logged, not retained, never used for training. |
Specifications
| Model | MiMo-V2.6-Flash by Xiaomi: 310B total, 12B active (mixture of experts), MIT licence |
|---|---|
| API model name | mimo-v2.6-flash |
| Context window | 262,144 tokens |
| Quantization | Native: fp8 attention, MXFP4 experts |
| Capabilities | Chat completions · streaming · tool calls · reasoning output when you ask for it · text only |
| Hardware | Two linked NVIDIA DGX Spark (GB10), owned and operated by AxForge |
| Region | Stockholm, Sweden (eu-se-1) |
| Input price | €0.33 / 1M tokens |
| Output price | €2.08 / 1M tokens |
Full details, benchmarks and FAQ on the MiMo-V2.6-Flash page. List prices in EUR.
How it works
| 1 | Sign in to AxForge. Your API key is created in the console. |
|---|---|
| 2 | Set base_url to https://api.axforge.ai/v1 in the OpenAI SDK you already use. |
| 3 | Send your first request with model mimo-v2.6-flash. |
| 4 | Watch usage in the console; every new account starts with free tokens. |
Try MiMo-V2.6-Flash or sign in to start.
FAQ
Yes. Chat completions, streaming, tool calls and usage reporting follow the OpenAI API shape, so the official SDKs work by changing base_url.
When you ask it to: pass chat_template_kwargs with "enable_thinking": true and the reasoning comes in reasoning_content, counted as output tokens. Without it, MiMo answers directly.
No. It reads text only, and a request with an image is refused with a clear message.
€0.33 per million input tokens and €2.08 per million output tokens. No subscription, no minimum.
In Stockholm, Sweden (eu-se-1), on NVIDIA DGX Spark systems AxForge owns and operates. TLS terminates in the EU.
No. Prompts and completions are processed in memory and never persisted or used for training.