The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30] WorkBuddy offers access inside a desktop agent; TokenHub provides hosted API access, while the open so...
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What did Xiaomi’s October 20, 2026 announcement of MiMo-V2.6-Pro’s availability in Tencent’s WorkBuddy desktop AI agent and Tencent Cloud To. Article summary: The October 20 announcement was primarily a distribution expansion: MiMo-V2.6-Pro became accessible through Tencent’s WorkBuddy desktop agent and TokenHub, rather than being introduced as a new model generation.[52] For . Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
Xiaomi’s October 20, 2026 announcement put MiMo-V2.6-Pro inside Tencent’s WorkBuddy desktop AI agent. The key distinction: this was a new route to use the model, not the model’s first release or first appearance on Tencent Cloud TokenHub. Xiaomi had released and open-sourced the Pro and Flash models on September 22, the same date Tencent’s product changelog says TokenHub added both models.11
14
30
These options differ in how the model is accessed and integrated. WorkBuddy is an application route, TokenHub is hosted inference, and the open-source release offers a separate route for teams that want to manage deployment themselves. Availability in an app or API catalog does not by itself mean usage is free, locally run, or identical across products.
MiMo-V2.6-Pro is listed with a 1-million-token context window, a maximum input of 1 million tokens, and a maximum output of 128,000 tokens. Tencent lists the same limits for Flash.20 These are published limits, not a guarantee that the maximum input and output can be used simultaneously in every request.
October launch coverage describes multi-agent task splitting and joint understanding of image, video, audio, and text, and positions the model for professional, long-running, complex workflows.11
13 Xiaomi’s own release materials describe Pro as intended for complex projects, long-horizon tasks, high-stakes work, cybersecurity, and research.
1 These are intended use cases, not guarantees of performance or suitability for consequential decisions.
Tencent’s TokenHub catalog also lists deep reasoning, structured output, function calling, caching, and multimodal understanding for Pro and Flash.20 The published TokenHub model guide describes text, image, and video input with text output for Pro; that specification should not be taken as confirmation that every modality or feature works identically in every client.
24
Tencent documents support for OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages protocols. Its integration guide says developers can use compatible SDKs, which can reduce the need to implement a different API protocol from scratch.23
25
That does not guarantee a drop-in replacement for every application. Developers still need to configure Tencent’s endpoint, credentials, and model ID, and should test the parameters, tool calls, and modality handling their software relies on.
Xiaomi says MiMo-V2.6-Pro scored 46 on the Artificial Analysis Intelligence Index, surpassing Kimi K3 and Qwen3.8 Max and taking the top open-source position at release.3 Treat that as Xiaomi’s reported claim: the provided sources do not include the underlying leaderboard or evaluation details needed to independently confirm the ranking. A score of 46 is also not a 46% accuracy rate, and it does not establish performance on every professional task.
Tencent’s international pricing page lists MiMo-V2.6-Pro at $0.435 per million uncached input tokens, $0.87 per million output tokens, and $0.0036 per million cached input tokens.22 Xiaomi’s model page lists the same USD rates for Pro.
15 Xiaomi’s domestic pay-as-you-go API pricing is listed separately at ¥3 per million uncached input tokens, ¥6 per million output tokens, and ¥0.025 per million cached input tokens.
2
Matching listed rates do not make the services’ billing arrangements interchangeable. Xiaomi says its self-serve API uses an ordinary API key and account balance, separate from its Token Plan quota.2 TokenHub pricing does not establish what WorkBuddy users pay, and app-level pricing should be checked separately.
The October news is best understood as a distribution update: it brought MiMo-V2.6-Pro into a third-party desktop agent, while cloud API access and the open-source release were already available through separate routes. That broadens where users can encounter the model, but the available reporting does not establish adoption numbers, business outcomes, or a new model architecture.11
14
30
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30]
The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30] WorkBuddy offers access inside a desktop agent; TokenHub provides hosted API access, while the open source release is a separate option for teams managing their own deployment.
Xiaomi reported a score of 46 on the Artificial Analysis Intelligence Index and claimed a top open source ranking; the supplied sources do not independently verify that ranking.[3]
The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30] WorkBuddy offers access inside a desktop agent; TokenHub provides hosted API access, while the open so...
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What did Xiaomi’s October 20, 2026 announcement of MiMo-V2.6-Pro’s availability in Tencent’s WorkBuddy desktop AI agent and Tencent Cloud To. Article summary: The October 20 announcement was primarily a distribution expansion: MiMo-V2.6-Pro became accessible through Tencent’s WorkBuddy desktop agent and TokenHub, rather than being introduced as a new model generation.[52] For . Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fa
Xiaomi’s October 20, 2026 announcement put MiMo-V2.6-Pro inside Tencent’s WorkBuddy desktop AI agent. The key distinction: this was a new route to use the model, not the model’s first release or first appearance on Tencent Cloud TokenHub. Xiaomi had released and open-sourced the Pro and Flash models on September 22, the same date Tencent’s product changelog says TokenHub added both models.11
14
30
These options differ in how the model is accessed and integrated. WorkBuddy is an application route, TokenHub is hosted inference, and the open-source release offers a separate route for teams that want to manage deployment themselves. Availability in an app or API catalog does not by itself mean usage is free, locally run, or identical across products.
MiMo-V2.6-Pro is listed with a 1-million-token context window, a maximum input of 1 million tokens, and a maximum output of 128,000 tokens. Tencent lists the same limits for Flash.20 These are published limits, not a guarantee that the maximum input and output can be used simultaneously in every request.
October launch coverage describes multi-agent task splitting and joint understanding of image, video, audio, and text, and positions the model for professional, long-running, complex workflows.11
13 Xiaomi’s own release materials describe Pro as intended for complex projects, long-horizon tasks, high-stakes work, cybersecurity, and research.
1 These are intended use cases, not guarantees of performance or suitability for consequential decisions.
Tencent’s TokenHub catalog also lists deep reasoning, structured output, function calling, caching, and multimodal understanding for Pro and Flash.20 The published TokenHub model guide describes text, image, and video input with text output for Pro; that specification should not be taken as confirmation that every modality or feature works identically in every client.
24
Tencent documents support for OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages protocols. Its integration guide says developers can use compatible SDKs, which can reduce the need to implement a different API protocol from scratch.23
25
That does not guarantee a drop-in replacement for every application. Developers still need to configure Tencent’s endpoint, credentials, and model ID, and should test the parameters, tool calls, and modality handling their software relies on.
Xiaomi says MiMo-V2.6-Pro scored 46 on the Artificial Analysis Intelligence Index, surpassing Kimi K3 and Qwen3.8 Max and taking the top open-source position at release.3 Treat that as Xiaomi’s reported claim: the provided sources do not include the underlying leaderboard or evaluation details needed to independently confirm the ranking. A score of 46 is also not a 46% accuracy rate, and it does not establish performance on every professional task.
Tencent’s international pricing page lists MiMo-V2.6-Pro at $0.435 per million uncached input tokens, $0.87 per million output tokens, and $0.0036 per million cached input tokens.22 Xiaomi’s model page lists the same USD rates for Pro.
15 Xiaomi’s domestic pay-as-you-go API pricing is listed separately at ¥3 per million uncached input tokens, ¥6 per million output tokens, and ¥0.025 per million cached input tokens.
2
Matching listed rates do not make the services’ billing arrangements interchangeable. Xiaomi says its self-serve API uses an ordinary API key and account balance, separate from its Token Plan quota.2 TokenHub pricing does not establish what WorkBuddy users pay, and app-level pricing should be checked separately.
The October news is best understood as a distribution update: it brought MiMo-V2.6-Pro into a third-party desktop agent, while cloud API access and the open-source release were already available through separate routes. That broadens where users can encounter the model, but the available reporting does not establish adoption numbers, business outcomes, or a new model architecture.11
14
30
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30]
The October 20, 2026 announcement was chiefly a WorkBuddy integration, not a new model release: Xiaomi had open sourced MiMo V2.6 Pro and Flash, and Tencent had listed them on TokenHub, on September 22.[11][14][30] WorkBuddy offers access inside a desktop agent; TokenHub provides hosted API access, while the open source release is a separate option for teams managing their own deployment.
Xiaomi reported a score of 46 on the Artificial Analysis Intelligence Index and claimed a top open source ranking; the supplied sources do not independently verify that ranking.[3]