MiMo V2.6 Pro Review: 1M Tokens, Four Input Types, a None Setting

On 28 October 2026 our platform sent MiMo V2.6 Pro one real request as part of launch: a single neutral user message, answered without streaming at low reasoning effort with a 1,024-token cap. It returned a complete answer, and the request was priced at the published rate. The endpoint also passed our pre-launch checks of its price table and request limits. During integration, media requests were accepted and answered.

That is a working-endpoint check, not a benchmark. We have not compared its answers with MiMo V2.6 Flash or another model, tested video and audio on real clips, measured speed, or tried a prompt near the 1,048,576-token limit.

These starting prompts come from the model page's published examples. We have not graded the outputs, so they are a place to begin rather than prompts proven on this model.

1. Critique an experiment plan. Give it the plan and ask for what is missing.

Review an experiment plan and identify assumptions, missing controls and reproducibility checks. Work from the material I provide, cite the relevant evidence, distinguish facts from assumptions, and identify missing information before proposing changes.

Run this prompt on MiMo V2.6 Pro →

2. A tool step. Confirms your function-calling loop end to end with a trivial function.

Call report_status with status ready.

Settings that matter: keep high for hard problems, drop to medium or low for easier ones, and use none when you want a direct answer with adjustable temperature; return reasoning_content on assistant turns when you continue a tool loop; keep inline video and audio under 14 MB combined.

We have not run our content-level test on MiMo V2.6 Pro, so we have no result to report here. The model page states the position plainly: content decisions come from the model, behaviour can vary, and no guarantee of unrestricted output is made.

Not tested yet: the content-level ladder, multi-turn behaviour over a long session, languages other than English, audio and video understanding on real material, and the none setting against the reasoning levels.

Every character in a request should be an adult, keep real, identifiable people out of it, and you are responsible for the law where you publish.

When the V2.6 family launched, one reader praised how much Xiaomi disclosed: "I know we have strong views on what a truly open model is (open weights, open training data, open training code etc.) but I really like how transparent they’ve been about the training of this model." Another summed up the pitch for the family: "This looks great in terms of cost and capabilities, truly pushing the frontier forward in terms of open weight light weight models." Both comments are about the family, not about our endpoint.

• The cheaper sibling? MiMo V2.6 Flash has the same window, reply ceiling and input types at a lower per-token price, and allows images up to 16,384 tokens.
• Always-on thinking and a million-token reply ceiling? Kimi K3.1 Pro takes images and video but not audio.
• Another 1M-class reasoning model? DeepSeek V4 Pro is on the same API; compare it on your own prompts.