Observed Signal · Sep 22, 2026 · Product Launch · Source: Trending Topics (DACH/CEE Innovation & Tech) · Impact: 4/5 · Sentiment: Positive
Xiaomi MiMo-V2.6-Pro tops open-weight AI models
Chinese electronics giant Xiaomi released its MiMo-V2.6 series of open-weight AI models under the MIT license on Hugging Face. The flagship MiMo-V2.6-Pro scored 46 points on the Artificial Analysis Intelligence Index, making it the highest-ranked open-weight model globally, surpassing GLM-5.3 and Kimi K3. It trails only proprietary models like Claude and GPT-6, ranking sixth overall. The model features a sparse mixture-of-experts architecture with 1.02 trillion total parameters (42 billion active), supports text, image, speech, and video input, and offers a one-million-token context window. Xiaomi trained the models using scaled reinforcement learning, live-streaming the production run and releasing weights, technical reports, and training code. The series also includes MiMo-V2.6-Flash and a faster UltraSpeed variant. API pricing remains unchanged from the previous generation.
A major tech company (Xiaomi) releases a top-performing open-weight AI model, potentially disrupting the AI model market and impacting AI-driven ad tech capabilities.
MARKET SIGNALS
Track Xiaomi Signals & Strategic Shifts
Polaris7 connects regulatory filings, primary sources, leadership changes, and M&A activity into structured intelligence. Set up your Polaris7 workspace to monitor relevant companies in your executive briefing.
- Verified primary sources
- Automated briefings
- Instant watchlist setup
Key Takeaways & Evidence Grounding
- Xiaomi released MiMo-V2.6-Pro, the top open-weight AI model with 46 points on the Artificial Analysis Intelligence Index.
- The model has 1.02 trillion parameters (42 billion active), a 1 million token context window, and multimodal input.
- Xiaomi trained the models using reinforcement learning over 750,000 trajectories in under six days.
- MiMo-V2.6-Pro API pricing is $0.435 per million input tokens and $0.87 per million output tokens.
- The model is available via Xiaomi's AI Studio, MiMo Code, MiMo Desktop, its API platform, and OpenRouter.
Connected Companies & Entities
“The Chinese electronics group Xiaomi has released the MiMo-V2.6 series of AI models...”
“The weights were placed under MIT license on Hugging Face.”
“The top model MiMo-V2.6-Pro scores 46 points on the Artificial Analysis Intelligence Index.”
“It replaces GLM-5.3 from Z.ai.”
“It also surpasses Kimi K3 from Moonshot AI.”
“Claude Fable 5.1 from Anthropic achieved 53 points.”
“GPT-6 Astra from OpenAI achieved 53 points.”
“Muse Spark 1.3 from Meta achieved 48 points.”
“It also outperformed DeepSeek V4.1 Pro, which scored 36 points.”
“Recently, Alibaba's Qwen3.8-Max has also pushed in this direction.”
“The model is also accessible via OpenRouter.”
“It outperformed Google's Gemini 3.8 Flash, which scored 41 points.”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Kimi K3.1 Pro: Open-Weight Model Beats GPT-6, Gemini 4
Chinese AI startup Moonshot AI unveiled Kimi K3.1 Pro, a new open-weight model that rivals leading US models in coding, 3D modeling, and financial analysis. It scored 55 on Artificial Analysis' Intelligence Index, ranking fourth overall and surpassing OpenAI's GPT-6 Astra and Google's Gemini 4 Argon (both 53). With weights to be released by November 4, it would become the strongest open-weight model, narrowing the gap to proprietary leaders from 12 to 3 points, trailing only Anthropic's Claude Opus 5.5 (58), Claude Sonnet 5.5 (56), and Meta's Muse Wonder Full (56). The model excels in agentic tasks, achieving top scores on Terminal-Bench 4.0 (67%, up from 13% for its predecessor) and 73% on AutomationBench-AA. API pricing is about one-third lower than Kimi K3 at $2 per million input tokens, $10 per million output tokens, and $0.20 for cached tokens. Notably, Moonshot removed cyberattack-related training data to reduce offensive capabilities. In other China AI news, a US Homeland Security order requires federal agencies to remove DeepSeek products; Wuxiang Cloud Valley signed a one-year AI token contract with a $200M annual minimum; MiniMax cut API prices by over 50% for M3.1 Flash; and Zhipu's Tmall store sold about 300 token units since September.
Xiaomi Uses IoT Margins to Fund AI Push
Xiaomi cut prices for its MiMo large language model API after a technical optimization reduced caching costs: a hierarchical KV cache increased cached token capacity fivefold and lowered caching costs by about 80 percent, enabling headline reductions of up to 99% for cache‑hit input tokens and broader input/output cuts of 60–80%. MiMo distributed 100 trillion free tokens to developers and briefly topped OpenRouter’s Hermes agent rankings before falling to 16th. The pricing move aligns MiMo with DeepSeek’s floor. Xiaomi’s Q1 2026 results provide financial context: revenue declined 10.9% year‑on‑year to RMB 99.1 billion and adjusted net profit fell 43.1% to RMB 6.1 billion; the electric-vehicle and AI segment posted a RMB 3.1 billion operating loss in Q1. Smartphone shipments, IoT revenue and EV deliveries showed mixed dynamics as management prioritized margin structure over headline growth.
Open-weight Models and Cost-Efficient AI Shift Market
Three major AI releases on July 28, 2026 — Moonshot AI's open-weight Kimi K3, Anthropic's Claude Opus 5, and Microsoft's MAI model family — signal a market shift from maximizing raw frontier performance toward cost-efficient, deployable models. Moonshot AI published full weights for Kimi K3 (a large Mixture-of-Experts model) enabling self-hosting and reduced vendor lock-in. Anthropic positioned Opus 5 as a lower-cost "daily driver" for most knowledge work, while Microsoft migrated core products to in-house MAI models claiming large GPU cost reductions versus OpenAI. Anthropic's CEO publicly argued against bans on open-weight models. The article frames winners as those delivering best performance-per-dollar across varied workloads.
WATCHLIST & BRIEFINGS
Receive future signals for Xiaomi & competitors in your briefing
Stay ahead of material M&A deals, leadership changes, account wins, and product rollouts. Polaris7 connects primary filings, company newsrooms, and trade reporting directly into your executive briefing.
- Regular primary-source checks
- Structured Watchlist briefings
- Preserves current context
Market Signal Monitoring
Follow affected companies in Polaris7 with regular source checks and structured Watchlist briefings.
Monitor Xiaomi
Follow affected companies in Polaris7 with regular source checks and structured Watchlist briefings.
