MiMo V2.6 Pro Review: 1M Tokens, Four Input Types, a None Setting
Pick MiMo V2.6 Pro for multimodal research and multi-step projects when you want audio and video in one request; so far we have run only launch checks.
Freedom Score
Not yet tested
- Price
- $0.544 per 1M mixed tokens (3 input : 1 output)
- Output speed
- 41 tok/s
- Success rate
- Collecting data
MiMo V2.6 Pro is the larger model in Xiaomi's V2.6 family, released as open weights in September 2026. Xiaomi positions it for complex multimodal projects and research. On our API it is one OpenAI-compatible chat endpoint that reads text, images, video and audio, holds 1,048,576 tokens of input and allows replies up to 131,072 tokens, reasoning included.
- SpicyAPI Engineering
- Tested Oct 28, 2026
- Methodology v3.3 →
- We sell these models; scores ignore price →
The verdict
Reasoning is on by default: reasoningeffort accepts none, minimal, low, medium or high, and defaults to high. Sampling has a catch. temperature (0 to 1.5) and topp apply only when reasoning_effort is none; while the model reasons it uses 1.0 and 0.95 and ignores your values.
This is a launch-day review. On 28 October 2026 our platform ran a real request through the endpoint as part of launch: one neutral message, answered without streaming at low reasoning effort. It returned a complete answer. We have not run a quality blind test or a content-level test, and this page does not score the model. Xiaomi publishes its own benchmark tables for the family; we have not reproduced them and do not quote them.
Key facts
- Tested: one real launch request on 28 Oct 2026 (neutral prompt, low reasoning effort, no streaming), plus integration checks of media input. No scored runs yet.
- Price: $0.435 per million input tokens, $0.0036 cached and $0.87 output. The live price is on the model page.
- Limits: 1,048,576-token context, replies up to 131,072 tokens (default 4,096, reasoning included), text, image, video and audio input, images up to 8,192 tokens each.
- Reasoning: reasoning_effort takes none, minimal, low, medium or high (default high).
- Sampling: temperature 0 to 1.5 and topp 0.01 to 1, applied only at reasoningeffort none.
- Sources: scoring rules are on our review methodology page, full parameters on the MiMo V2.6 Pro model page.
What we liked and what held it back
What we liked
- Four input types in one window. text, imageurl, videourl, audiourl and inputaudio (wav or mp3) parts can sit in one request, so a screen recording, a voice memo and the question travel together.
- A none setting. Set reasoningeffort to none for a direct answer, and temperature and topp become adjustable.
- A low per-token price for a Pro model. $0.435 input and $0.87 output per million tokens.
- A 131,072-token reply ceiling. Enough for a full report or a long extracted table in one reply.
- Tool calling is on the endpoint. Up to 128 function definitions with toolchoice and paralleltool_calls; we have not graded tool-use quality yet.
What held it back
- Reasoning is on by default, at high. Every request reasons unless you set none, which costs more output tokens than a lighter default.
- Sampling is conditional. temperature and top_p are ignored while the model reasons, so you only get control by turning reasoning off.
- Smaller image budget than Flash. Images are capped at 8,192 tokens each here, against 16,384 on MiMo V2.6 Flash, so downscale very large pages.
- Reasoning is paid for out of max_tokens. A small cap at high effort can be used up before the visible answer is complete; check finish_reason.
- No quality numbers of our own yet.
What we tested
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.
Prompts to start from
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.
How far it writes
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.
What the community says
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.
Alternatives
- 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.
How MiMo V2.6 Pro earns its Spicy Index
Each dimension below links to its full workings: the table row every point came from and why points were lost.
- Public leaderboards 40% of quality
- Not yet tested
- Role-play 15% of quality
- Not yet tested
- Creative writing 15% of quality
- 63.3
- Instruction following 15% of quality
- 99.0
- Long-context memory 15% of quality
- Not yet tested
- Capability 10.5% weight
- 64.0
- Freedom Score 30% weight
- Not yet tested
No exact match on a tracked public leaderboard yet. How it works
Our role-play suite hasn’t run on this model yet. How it works
Our long-context suite hasn’t run on this model yet. How it works
Freedom evidence is untested or insufficient. The Index retains the original quality and capability score, is marked preliminary and still receives a rank. Missing evidence is not zero. How it works
Freedom evidence is untested or insufficient. The Index retains the original quality and capability score, is marked preliminary and still receives a rank. Missing evidence is not zero.
Open the full test report →Capability, item by item
Points from the public scoring table. Upscaled tiers and limits we set ourselves don’t count.
- Context window (creator-published)Not declared
- 0.0 / 30
- Input modalities on SpicyAPIText, plus images, audio
- 19.0 / 25
- Tool useTools and tool_choice
- 15.0 / 15
- Structured outputAccepted, schema adherence verified
- 15.0 / 15
- Reasoning control5 levels: none · minimal · low · medium · high
- 15.0 / 15
What MiMo V2.6 Pro can do on SpicyAPI
Straight from the live model specification on SpicyAPI. Some creators offer more in their own apps; this is what you can call through our API today.
| Feature | Chat |
|---|---|
| Context window | 1,048,576 tokens |
| Max output per request | 131,072 tokens |
| Reasoning levels | none · minimal · low · medium · high |
| Tool calling | Yes |
| Structured output | Yes |
How long it really takes
Measured on real jobs over the last 30 days, from request to finished result. Times include our own test runs.
- Output speed
- 41 tok/sIncludes time to first token
- Success rate
- Collecting data
A time appears once at least 10 completed jobs are in. Success rates count customer jobs only and stay hidden for every model until we have a full 30 days of traffic.
What people say about MiMo V2.6 Pro
2 positive · 0 mixed · 0 negative
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.
hacker-news· Sep 21, 2026· PositiveSource
This looks great in terms of cost and capabilities, truly pushing the frontier forward in terms of open weight light weight models.
hacker-news· Sep 21, 2026· PositiveSource
Quotes are verbatim from public posts, linked to the original. They are not ratings and don’t feed any score.
Models to consider instead of MiMo V2.6 Pro
Xiaomi
MiMo V2.6 Flash
The cheaper sibling?
- Spicy Index
- 76.2−2.4
- Freedom Score
- —
- Price
- $0.175−68%/ 1M tokens · 3:1 input/output
Moonshot AI
Kimi K3.1 Pro
Always-on thinking and a million-token reply ceiling?
- Spicy Index
- 80.4+1.8
- Freedom Score
- —
- Price
- $4.00+636%/ 1M tokens · 3:1 input/output
DeepSeek
DeepSeek V4 Pro
Another 1M-class reasoning model?
- Spicy Index
- 79.0+0.4
- Freedom Score
- 88.6
- Price
- $1.98+264%/ 1M tokens · 3:1 input/output
What do you think?
Was this review useful?
Log in to vote.
Comments
Comments are checked automatically and by our team. We remove spam, offensive content, personal attacks and anything that names a real person without consent. Report a comment and we review it within one business day.
Tried this model? Share what worked. You need a free account to comment, so we can keep spam out.
MiMo V2.6 Pro FAQ
Try MiMo V2.6 Pro yourself
Set every parameter in the Playground and see the exact price before you run it.
$0.544 per 1M mixed tokens (3 input : 1 output)
More reviews
See the leaderboards →GPT 6.1 Terra
Pick GPT 6.1 Terra for high-volume screens, documents and writing where reasoning_effort none helps; so far we have run only launch checks, no scored tests.
- Spicy Index
- 79.9
- Freedom Score
- —
Tested Oct 28, 2026
Read the reviewKimi K3.1 Pro
Pick Kimi K3.1 Pro for long-context work with screenshots and clips when you accept always-on thinking; so far we have run only launch checks, no scored tests.
- Spicy Index
- 80.4
- Freedom Score
- —
Tested Oct 28, 2026
Read the review
Grok 4.3
Pick Grok 4.3 for a huge context window, a genuine reasoning off switch and uncensored writing at a low price; skip it if you need file uploads.
Tested Sep 24, 2026
Read the review