<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm" -->

---
title: “Impressive level of openness”: Xiaomi goes way beyond...
description: Xiaomi released MiMo-V2.6, a trillion-parameter open-weight model family, alongside an unusually transparent training process: it livestreamed its...
canonical: https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: “Impressive level of openness”: Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6 | daily.dev
og:description: Xiaomi released MiMo-V2.6, a trillion-parameter open-weight model family, alongside an unusually transparent training process: it livestreamed its...
og:url: https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm
og:image: https://api.daily.dev/og/posts/qmPOxZwpM.png
og:image:alt: “Impressive level of openness”: Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# “Impressive level of openness”: Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6

**[The New Stack](https://daily.dev/sources/newstack)** · 6 min read · 0 upvotes · 0 comments

## Summary

Xiaomi released MiMo-V2.6, a trillion-parameter open-weight model family, alongside an unusually transparent training process: it livestreamed its reinforcement-learning runs publicly for five days, disclosed exact RL training costs (~$3.5 million combined for the Flash and Pro variants), released weights under the permissive MIT license, and promised to open-source over 7,000 RL task environments and its training framework. Independent benchmarking from Artificial Analysis ranks MiMo-V2.6-Pro first among 114 tracked open-weight models. Researchers including Hugging Face's Thomas Wolf and Prime Intellect's Elie Bakouch praised the openness, especially the promised release of RL environments, which are increasingly valuable as training shifts toward reinforcement learning with verifiable rewards (RLVR). The piece also contrasts this approach with less permissive open-weight releases from Meta, Moonshot, and Alibaba.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://thenewstack.io/xiaomi-mimo-vs-6-open-source>

## Questions this post answers

### What is Xiaomi's MiMo-V2.6-Pro model and how big is it?

MiMo-V2.6-Pro is a trillion-parameter large language model from Xiaomi with 42 billion parameters active at a time, a one-million-token context window, and support for text, images, audio, and video. It's released under the permissive MIT license with no commercial restrictions, and Artificial Analysis ranks it first among 114 tracked open-weight models with an Intelligence Index score of 46.

_daily.dev surfaces releases like MiMo-V2.6 for developers comparing open-weight model options._

### How much did Xiaomi's MiMo-V2.6 reinforcement learning training cost?

The RL training phase cost about $3.5 million combined: $854,044 for the smaller MiMo-V2.6-Flash model and $2,620,670 for the larger Pro model, based on a public dashboard Xiaomi ran during a five-day livestreamed training run starting September 15. This figure covers only the RL stage, not pretraining costs. By comparison, MiniMax-M1's RL phase cost $534,700 and DeepSeek R1's cost about $294,000.

_track real training-cost disclosures like this on daily.dev when budgeting your own RL experiments._

### What is the difference between open-weight and open-source AI models?

Open-weight means only the trained numerical parameters are released for running or fine-tuning a model, while much of the training process, data, and code stays closed. Open-source implies the fuller pipeline, including training data, environments, and frameworks, is also made available. Meta's Llama has been criticized for calling itself open-source despite licensing restrictions, while Xiaomi's MiMo-V2.6 uses an unrestricted MIT license and plans to release its RL training environments and framework.

_daily.dev helps developers untangle open-weight versus open-source claims before adopting a model._

## Similar posts on daily.dev

- [Xiaomi releases MIT‑licensed MiMo models for long‑running AI agents](https://daily.dev/posts/xiaomi-releases-mit-licensed-mimo-models-for-long-running-ai-agents-qgbigdop8) · InfoWorld · 2 upvotes · 0 comments

---

Tags: [#tech-news](https://daily.dev/tags/tech-news), [#open-source](https://daily.dev/tags/open-source), [#llm](https://daily.dev/tags/llm), [#reinforcement-learning](https://daily.dev/tags/reinforcement-learning)

[View this post on daily.dev](https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"“Impressive level of openness”: Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6","url":"https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm"},"datePublished":"2026-09-23T20:04:54.878Z","dateModified":"2026-09-26T06:14:29.264Z","description":"Xiaomi released MiMo-V2.6, a trillion-parameter open-weight model family, alongside an unusually transparent training process: it livestreamed its...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/cf51f7061e31c3f554196bd3c8a54ee6?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/cf51f7061e31c3f554196bd3c8a54ee6?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"tech-news,open-source,llm,reinforcement-learning","timeRequired":"PT6M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"“Impressive level of openness”: Xiaomi goes way beyond the usual open-weight playbook with MiMo-V2.6"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/impressive-level-of-openness-xiaomi-goes-way-beyond-the-usual-open-weight-playbook-with-mimo-v2-6-qmpoxzwpm#faq","mainEntity":[{"@type":"Question","name":"What is Xiaomi's MiMo-V2.6-Pro model and how big is it?","acceptedAnswer":{"@type":"Answer","text":"MiMo-V2.6-Pro is a trillion-parameter large language model from Xiaomi with 42 billion parameters active at a time, a one-million-token context window, and support for text, images, audio, and video. It's released under the permissive MIT license with no commercial restrictions, and Artificial Analysis ranks it first among 114 tracked open-weight models with an Intelligence Index score of 46. daily.dev surfaces releases like MiMo-V2.6 for developers comparing open-weight model options."}},{"@type":"Question","name":"How much did Xiaomi's MiMo-V2.6 reinforcement learning training cost?","acceptedAnswer":{"@type":"Answer","text":"The RL training phase cost about $3.5 million combined: $854,044 for the smaller MiMo-V2.6-Flash model and $2,620,670 for the larger Pro model, based on a public dashboard Xiaomi ran during a five-day livestreamed training run starting September 15. This figure covers only the RL stage, not pretraining costs. By comparison, MiniMax-M1's RL phase cost $534,700 and DeepSeek R1's cost about $294,000. track real training-cost disclosures like this on daily.dev when budgeting your own RL experiments."}},{"@type":"Question","name":"What is the difference between open-weight and open-source AI models?","acceptedAnswer":{"@type":"Answer","text":"Open-weight means only the trained numerical parameters are released for running or fine-tuning a model, while much of the training process, data, and code stays closed. Open-source implies the fuller pipeline, including training data, environments, and frameworks, is also made available. Meta's Llama has been criticized for calling itself open-source despite licensing restrictions, while Xiaomi's MiMo-V2.6 uses an unrestricted MIT license and plans to release its RL training environments and framework. daily.dev helps developers untangle open-weight versus open-source claims before adopting a model."}}]}
```

