Xiaomi Files Trademarks for MIMO ORBIT and MIMO CLAW
Xiaomi's new trademark filings often hint at upcoming hardware or robotics projects with AI potential.
30-Second TL;DR
What Changed
Trademark filings for 'MIMO ORBIT' and 'MIMO CLAW'
Why It Matters
Trademark filings often precede new product launches, potentially indicating Xiaomi's expansion into robotics or advanced AI-integrated hardware.
What To Do Next
Keep an eye on Xiaomi's robotics and IoT product announcements for potential integration of these new brand names.
Key Points
- •Trademark filings for 'MIMO ORBIT' and 'MIMO CLAW'
- •Scope includes scientific instruments and web services
- •Status currently under substantive examination
Deep Insight
Background and context from public sources — not the original article. 10 sources cited.
Enhanced Key Takeaways
- •Xiaomi has recently open-sourced its MiMo-V2.5 series of large language models, including a native full-modal model and a version optimized for Agent and Coding applications, under the permissive MIT License.
- •The 'MiMo Orbit Program' has been launched by Xiaomi, featuring a 'Creator Trillion Token Incentive Plan' to distribute 100 trillion tokens to AI builders and an 'Agent Ecosystem Co-construction Plan' for Agent framework teams.
- •The term 'MiMo Claw' refers to the advanced agentic task capabilities of Xiaomi's MiMo models, with the MiMo-V2.5-Pro model demonstrating high performance in such tasks on the ClawEval leaderboard.
- •The MiMo-V2.5 series models are designed to support a 1-million-token context window and offer comprehensive understanding across text, image, video, and audio inputs.
Competitor Analysis
- Xiaomi MiMo-V2.5-Pro
- Optimal frontier for task completion rate and token efficiency
- Claude Opus 4.6
- 81.5
- Claude Sonnet 4.6
- 79.2
- GPT-5.2
- 77.0
- Gemini 3 Pro
- 67.7
- Kimi K2.6
- N/A
- GLM 5.1
- N/A
- Xiaomi MiMo-V2.5-Pro
- 1581
- Claude Opus 4.6
- N/A
- Claude Sonnet 4.6
- N/A
- GPT-5.2
- N/A
- Gemini 3 Pro
- N/A
- Kimi K2.6
- Lower than MiMo-V2.5-Pro
- GLM 5.1
- Lower than MiMo-V2.5-Pro
- Xiaomi MiMo-V2.5-Pro
- MIT License (Open Source)
- Claude Opus 4.6
- Proprietary
- Claude Sonnet 4.6
- Proprietary
- GPT-5.2
- Proprietary
- Gemini 3 Pro
- Proprietary
- Kimi K2.6
- Proprietary
- GLM 5.1
- Proprietary
- Xiaomi MiMo-V2.5-Pro
- 1-million-token
- Claude Opus 4.6
- N/A
- Claude Sonnet 4.6
- N/A
- GPT-5.2
- N/A
- Gemini 3 Pro
- N/A
- Kimi K2.6
- N/A
- GLM 5.1
- N/A
- Xiaomi MiMo-V2.5-Pro
- Full-modal (text, image, video, audio)
- Claude Opus 4.6
- N/A
- Claude Sonnet 4.6
- N/A
- GPT-5.2
- N/A
- Gemini 3 Pro
- N/A
- Kimi K2.6
- N/A
- GLM 5.1
- N/A
| Feature/Benchmark | Xiaomi MiMo-V2.5-Pro | Claude Opus 4.6 | Claude Sonnet 4.6 | GPT-5.2 | Gemini 3 Pro | Kimi K2.6 | GLM 5.1 |
|---|---|---|---|---|---|---|---|
| ClawEval (Agentic Tasks) | Optimal frontier for task completion rate and token efficiency | 81.5 | 79.2 | 77.0 | 67.7 | N/A | N/A |
| GDPVal-AA (Elo) | 1581 | N/A | N/A | N/A | N/A | Lower than MiMo-V2.5-Pro | Lower than MiMo-V2.5-Pro |
| License | MIT License (Open Source) | Proprietary | Proprietary | Proprietary | Proprietary | Proprietary | Proprietary |
| Context Window | 1-million-token | N/A | N/A | N/A | N/A | N/A | N/A |
| Modality | Full-modal (text, image, video, audio) | N/A | N/A | N/A | N/A | N/A | N/A |
Technical Deep Dive
- Model Series: MiMo-V2.5, comprising MiMo-V2.5 (native full-modal specialist) and MiMo-V2.5-Pro (Agent specialist).
- Optimization: MiMo-V2.5-Pro is specifically engineered for 'long-horizon coherence' and complex software engineering tasks.
- Context Window: Both models in the MiMo-V2.5 series support a 1-million-token context window.
- Modality: MiMo-V2.5 is a native full-modal model, capable of understanding text, image, video, and audio.
- Licensing: The MiMo-V2.5 series is released under the MIT License, permitting commercial inference deployment and secondary training without additional authorization.
- Performance: MiMo-V2.5-Pro achieved a score of 1581 on the GDPVal-AA (Elo) benchmark and ranks at the optimal frontier for task completion rate and token efficiency on the Claw-Eval leaderboard.
Future ImplicationsAI analysis grounded in cited sources
Timeline
- 2010Xiaomi Corporation founded
- 2019Xiaomi's Science and Technology Park in Beijing inaugurated, housing 128 laboratories, indicating a strong R&D focus
- 2023-10Xiaomi 'MIOS' trademarks appeared, classified in scientific instruments, communication services, and website services, showing early intent in these areas
- 2025-12-19Xiaomi completed a 5,000-square-meter Sports and Health Lab, demonstrating investment in scientific instruments for wearables
- 2026-02-25Xiaomi filed 'Xiaomi Intelligent Storage' trademarks, also covering scientific instruments and web/telecom services, further indicating a broad strategic push
- 2026-04-27Xiaomi open-sourced the MiMo-V2.5 series of AI models and launched the MiMo Orbit Program, directly linking 'MiMo Orbit' and 'MiMo Claw' to advanced AI capabilities and ecosystem development
Sources (10)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
Weekly AI Recap
Read this week's curated digest of top AI events →
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪 ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
The weekly digest
One email a week. Unsubscribe anytime.