AI Daily: Kimi K3.1 Pro Hits Microsoft Foundry, DeepSeek Harness DAU Tops 1.54M
快鲤鱼 · wechat · 2026-11-06
A Chinese AI industry digest rounding up multiple items:
- Kimi K3.1 Pro launches on Microsoft Foundry: Moonshot AI's open-weight agentic model with vision-language understanding and up to 1M token context — Microsoft cloud's first revenue-share open-weight model, its second major cloud partner after AWS
- DeepSeek Harness DAU passes 1.5M: reached 1.54M daily actives and 92T tokens/day on Nov 4, up from 271K on Sept 30
- Zhang Chaoyang on LLMs: says Sohu "can't afford" frontier models after missing the mobile era, sees opportunity in overlooked niches
- World Labs ships Frameboy: Fei-Fei Li's team brings its autoregressive world model Atlas to real-time interactive roaming, integrated into the closed-beta Viewfinder web app — first deliverable since joining AMD
- OpenAI launches GPT-6 Cyber: a security model that finds and validates exploitable vulnerabilities within researcher-set scopes, outperforming GPT-5.6 Cyber and GPT-6 Astra; initially limited to vetted Daybreak Red customers
- Meta Muse adds real-time voice: interruptible conversations plus avatars streaming 448×768@25fps portrait video with 870ms first-response latency
Related event: Kimi K3.1 Pro Lands on Microsoft Foundry(2 posts)→
More from Models
- ChatGPT reportedly mass-banning accounts, long-time paying users included — zyyyyl9144 · 2026-11-07
- Dev runs 284B-parameter DeepSeek-V4-Flash on CPU only, squeezed into 64GB DDR5 RAM — solyarisoftware · 2026-11-07
- Inkling 2 Small jumps +240 pts in Code Arena: WebDev, would be #1 US open model — SonglinYang4 · 2026-11-07
- MIT's Vector Policy Optimization replaces scalar RL rewards to boost LLM test-time search, accepted at NeurIPS 2026 — kastnerkyle · 2026-11-07
- Uncensored DeepSeek-distilled Qwen3.5-9B merge hits HF trending — xtools-at · 2026-11-07
- Anthropic paper: 3M tokens of moral dilemmas cut model malicious behavior to 0%, 28x efficiency — garrytan · 2026-11-07