MiMo-V2.6-Pro
MiMo-V2.6-Pro is a natively omnimodal sparse Mixture-of-Experts model with 1.02T total and 42B activated parameters and a 1M-token context, which Xiaomi calls its most capable MiMo model. Xiaomi released and open-sourced it under the MIT licence on September 22, 2026.
Released
September 22, 2026
Type
multimodal
License
open-weight
MIT
Context
—
Xiaomi states a context window of “1M” tokens.
Pricing
Input
$0.435 / 1M tokens
Output
$0.87 / 1M tokens
List price on the developer's own pages, checked . Overseas real-time API, cache-miss input. Cache-hit input $0.0036; batch $0.2175 / $0.435. The faster MiMo-V2.6-Pro-UltraSpeed is $4.35 / $8.70. In China it is ¥3 / ¥6.
Capabilities
Architecture
Parameters: 1.02T total, 42B activated (sparse MoE); 681M-parameter vision encoder
Sources
Every figure on this page was checked against these pages on . A figure we could not source is left out.
- Xiaomi MiMo: Introducing MiMo-V2.6 series (September 22nd, 2026) (opens in a new tab)
- Xiaomi MiMo API: Pay-as-you-go pricing (opens in a new tab)
- Hugging Face: XiaomiMiMo/MiMo-V2.6-Pro-RL (opens in a new tab)
- Hugging Face: XiaomiMiMo/MiMo-V2.6-Pro-MOPD (opens in a new tab)
Spotted something out of date? Tell us.
Links
In the news: MiMo-V2.6-Pro and Xiaomi
9to5Google
Arm’s new CPU cores and GPUs for Android chips promise ‘desktop-class’ games, more AI power
Arm Holdings, a leading designer of semiconductor intellectual property (IP), has unveiled its next-generation CPU cores, codenamed "C2," and significant upgrades to its Mali Graphics Processing Units (GPUs). These advancements are specifically engineered for future System-on-Chips (SoCs) destined for Android devices, with a clear ambition to elevate mobile gaming to "desktop-class" levels and substantially enhance on-device artificial intelligence (AI) processing power. Arm's IP is foundational to the vast majority of mobile processors, making these announcements critical for the entire Android ecosystem. The C2 CPU cores represent a substantial leap forward in performance and efficiency compared to their predecessors. Designed to be integrated into the central processing units of mobile SoCs, these cores are optimized for sustained performance, a crucial factor for demanding applications like high-fidelity mobile gaming and intensive multitasking. This focus on sustained power delivery aims to prevent performance throttling during extended use, ensuring a smoother and more responsive user experience. Concurrently, the updated Mali GPUs are being introduced to deliver a dramatic improvement in graphical rendering capabilities. These new GPU architectures are tailored to render more complex visual effects, higher polygon counts, and improved lighting, bringing the visual fidelity of mobile games closer to that of dedicated gaming consoles and PCs. Arm has also emphasized increased power efficiency in these GPU upgrades, allowing for these enhanced graphical experiences without a disproportionate drain on battery life, a perennial concern for mobile device users. Beyond the realm of gaming, Arm is placing a significant emphasis on the augmented AI capabilities of its new chip designs. Both the C2 CPU cores and the upgraded Mali GPUs are being optimized to accelerate a wide range of AI workloads. This includes on-device machine learning inference, which allows AI models to run directly on the device rather than relying on cloud servers. Such local processing offers benefits in terms of speed, privacy, and reduced data usage. Specific AI applications that will see improvements include advanced image processing for cameras, real-time natural language understanding for voice assistants and translation, and the delivery of more personalized user experiences through intelligent software. The growing reliance of applications on AI features makes this enhancement a key strategic priority for Arm. These new core designs are not sold directly to consumers but are licensed to major semiconductor manufacturers, including industry giants like Qualcomm, MediaTek, and Samsung. These companies then integrate Arm's IP into their own proprietary SoCs, which are subsequently supplied to smartphone and tablet manufacturers such as Samsung, Google, Xiaomi, and others. While Arm has not provided a specific timeline for the release of devices featuring these new C2 cores and Mali GPUs, the typical product development cycle suggests their integration into flagship Android devices within the next 12-18 months. This strategic evolution of its core IP reinforces Arm's dominant position in the mobile processor market, ensuring that the foundational technology powering the world's smartphones and tablets continues to advance.
Electrek
Xiaomi delivers on Tesla’s decade-old robot charger vision with new home robotic arm
Xiaomi unveiled a home robotic arm charger on October 18, 2024, that autonomously plugs and unplugs electric vehicles. This device fulfills a concept that Tesla prototyped more than ten years ago but did not commercialize. The charger is 152 mm wide, designed for narrow home garages, and connects to Xiaomi's smart home system for control via smartphone. The robotic arm can extend up to 1.5 meters and is capable of identifying the charging port with its integrated camera and sensors. It is designed to handle various charging connectors, including Type 1 and Type 2 standards. Xiaomi stated that the charger is compatible with their SU7 electric vehicle and aims to provide a seamless charging experience for users. The company has not yet announced pricing or availability details for the charger, but it represents a significant step in home EV charging automation. This innovation addresses a common inconvenience for EV owners, particularly those with limited mobility or in challenging parking situations. The integration with Xiaomi's existing smart home platform suggests a future where home appliances and vehicles operate in a more connected and automated manner. The robotic arm's ability to precisely locate and connect to the charging port highlights advancements in robotic vision and manipulation technology.