MiMo-V2.5
Xiaomi · released Apr 27, 2026 · XiaomiMiMo/MiMo-V2.5
- Type
- Open weightsMIT License
- Params
- 311B
- Context
- 1.1M
active per word not recorded by us · about 788K words of context
Our take
Written Sep 2, 2026MiMo-V2.5 is a large downloadable model from Xiaomi with a permissive MIT licence and a one-million-token request limit. It scores highest on coding tasks among its measured variants, and handles text, images, audio and video input.
Pick this for open-weights coding workloads where a permissive licence matters, or for very long-context tasks up to 1.05 million tokens. Use it if you want provider choice at competitive rates, with fourteen tracked offers. Skip it if creative writing quality is central, or if you need verified efficiency claims from disclosed active parameters.
The case for it
- Strongest measured skill is coding, with web-development coding also above its overall average.
- Hard prompts and mathematics both outperform its overall text rating.
- Permissive MIT licence allows proprietary use and redistribution without attribution.
- One-million-token request limit stands among the largest in downloadable models.
The case against it
- Creative writing is the lowest of its seven measured variants, nearly forty points below its overall score.
- No disclosed active-parameter count, leaving efficiency claims unverifiable.
- The vendor's own endpoint is the slowest measured, at less than a third of the fastest provider's throughput.
How good is it?
EverydayGeneral questions and everyday reasoning
Arena Text (overall)67th of 168 · 1434
CodingWriting and fixing code on its own
Arena Coding60th of 168 · 1491
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing74th of 168 · 1394
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model7 scoresEvery figure we hold, from 7 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Comfortable fit
Apple M3 Ultra (80-core GPU) · 512 GB
Room to spare. 180.7 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked between 60 min and 36 days ago — each listing carries its own date.
The only listing at 1.1M of context — the other 7 in the table below are not like-for-like. One cheaper row there is outside that comparison: a different quantisation.
- per 1M tokens
- $0.14 in / $0.28 out
- Context served
- 1.1M
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| GMICloudfp8Through OpenRouter | $0.12 / $0.24checked 2 days ago | 1.1M945K max reply | 25 tok/s | No | Yesunknown period | Unknown |
| DeepInfrafp8Direct | $0.14 / $0.28checked 9 days ago | 262K | not measured | Unknown | Unknown | Unknown |
| OpenRouterOpenRouter's own listing | $0.14 / $0.28checked 1 hour ago | 1.1M | not measured | Unknown | Unknown | Unknown |
| Xiaomifp8Through OpenRouter | $0.14 / $0.28checked 60 min ago | 1M131K max reply | 40 tok/s | No | Yes30 days | Unknown |
| Novita AIfp8Direct and through OpenRouter | $0.17 / $0.34checked 1 hour ago directchecked 60 min ago through OpenRouter | 1M131K max reply through OpenRouter | 38 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
| StreamLakeThrough OpenRouter | $0.17 / $0.34checked 60 min ago | 1M128K max reply | 33 tok/s | No | Yesunknown period | Unknown |
| DeepInfraDirect | $0.40 / $2.00checked 36 days ago | 262K | not measured | Unknown | Unknown | Unknown |
| Venice AIfp8Through OpenRouter | $0.40 / $2.00checked 60 min ago | 1M66K max reply | 21 tok/s | No | No | Confirmed |
DeepInfra
Novita AIAcross the 8 listings we hold: 5 say they do not train on prompts (1 of them only through OpenRouter), 0 say they do and 3 do not say. 2 appear in the zero-retention registry we check (1 of them only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| GMICloudfp8Through OpenRouter | ✓ | ✗ | ✗ |
| DeepInfrafp8Direct | |||
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Xiaomifp8Through OpenRouter | ✓ | ✓ | ✗ |
| Novita AIfp8Direct and through OpenRouter | ✓ | ✓ | ✗ |
| StreamLakeThrough OpenRouter | ✗ | ✓ | ✓ |
| DeepInfraDirect | |||
| Venice AIfp8Through OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 5 of 8 listings say yes, 1 says no, 2 publish no parameter list. JSON output: 5 of 8 listings say yes, 1 says no, 2 publish no parameter list. Strict schema: 3 of 8 listings say yes, 3 say no, 2 publish no parameter list.
Models people weigh against MiMo-V2.5
When we formed this view
Recent changes
What moved
input −65% ($0.40 → $0.14 per 1M tokens), output −86% ($2.00 → $0.28 per 1M tokens)What moved
input +61% ($0.130 → $0.209 per 1M tokens), output +23% ($0.26 → $0.32 per 1M tokens), cache read +61% ($0.0650 → $0.1045 per 1M tokens)What moved
input −23% ($0.168 → $0.130 per 1M tokens), output −23% ($0.336 → $0.260 per 1M tokens)Each date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 2 of 8 listings publish no parameter list, so what their API accepts is unknown to us.
- We do not hold the active parameter count for it, so how much of it runs on any one token is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 3 of 8 listings do not say whether they train on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsMIT License, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
MIT License
Fully permissive: do anything with attribution. No patent grant, unlike Apache-2.0.
Identifiers
- Hugging Face
- XiaomiMiMo/MiMo-V2.5
- Architecture
- Mixture of experts
- Takes in, gives back
- Text, images, audio and video in, text out
- Catalogue slug
- xiaomi-mimo-v2-5