MiMo-V2.6-Flash on QuickSilver Pro
MiMo-V2.6-Flash is the fast, low-cost sibling of Xiaomi's MiMo-V2.6-Pro — a 309B-parameter mixture-of-experts with 15B active. It accepts text and images across a 1M-token context with streaming, tools, and structured output. QuickSilver Pro charges $0.112 input / $0.224 output per million tokens — 20% below Xiaomi's $0.14 / $0.28 list price, and the same price as MiMo-V2.5.
At a glance
Fast, low-cost open-weight coding and agents with image input and a 1M-token context.
Pricing comparison ($/1M tokens)
| Provider | Input | Output | vs QSP |
|---|---|---|---|
| QuickSilver Pro | $0.112 | $0.224 | lowest-cost |
| Xiaomi list price (Xiaomi MiMo API) | $0.14 | $0.28 | 20% lower |
When to use
Use MiMo-V2.6-Flash for high-volume coding loops, tool-calling agents, extraction and classification, batch codegen, and screenshot or document understanding where cost and latency matter more than peak reasoning. It is the natural upgrade from MiMo-V2.5: same per-token price, newer generation, and image input.
When to use something else
For the hardest coding, multi-file refactors, and long agent chains, step up to MiMo-V2.6-Pro ($0.348/$0.696), the top open-weights model on the Artificial Analysis Intelligence Index. For pure mathematical reasoning, DeepSeek V4 Pro. MiMo-V2.6-Flash runs in non-thinking mode on QuickSilver Pro.
Quickstart (curl)
curl https://api.quicksilverpro.io/v1/chat/completions \
-H "Authorization: Bearer $QSP_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "mimo-v2.6-flash",
"messages": [{"role": "user", "content": "Hello!"}]
}'OpenAI-compatible. One-line migration via base_url.
FAQ
MiMo-V2.6-Flash is Xiaomi's newer-generation fast model: a 309B / 15B-active MoE that adds image input, at exactly the same QuickSilver Pro price as MiMo-V2.5 ($0.112 / $0.224). MiMo-V2.5 remains available, so switching is a one-line model change.
No. QuickSilver Pro serves it in non-thinking mode, so calls return a direct answer without billing a hidden reasoning trace. A `reasoning` field in the request does not turn reasoning on today.
QuickSilver Pro is $0.112 input / $0.224 output per million tokens, and $0.00224 per million cached-input tokens — 20% below Xiaomi's list price of $0.14 / $0.28 ($0.0028 cached), lower on both legs.
Yes. Set base_url=https://api.quicksilverpro.io/v1, paste your QSP key, and use model="mimo-v2.6-flash". Chat Completions and the Responses API both work, with streaming, tool calling, json_schema strict mode, image input, and usage.cost accounting.