SenseTime

🔥 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮 𝗨𝟭.𝟱 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗥𝗲𝗽𝗼𝗿𝘁 𝗶𝘀 𝗼𝘂𝘁. From capability coexistence to 𝗴𝗲𝗻𝘂𝗶𝗻𝗲 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗶𝗻𝘁𝗲𝗿𝗮𝗰𝘁𝗶𝗼𝗻. An open-source 8B-MoT native unified model that goes beyond "seeing + drawing." Inside: ✨ 𝗡𝗮𝘁𝗶𝘃𝗲 𝗠𝗼𝗧 — understanding & generation connected through shared attention ✨ 𝗦𝗽𝗮𝘁𝗶𝗮𝗹𝗹𝘆 𝗝𝗼𝗶𝗻𝘁 𝗥𝗲𝗰𝗼𝗻𝘀𝘁𝗿𝘂𝗰𝘁𝗶𝗼𝗻 — Pixel Shuffle + 3×3 spatial convolutions for native 4K generation and stronger spatial coherence ✨ 𝗦𝗽𝗲𝗰𝗶𝗮𝗹𝗶𝘇𝗲 ➡️ 𝗨𝗻𝗶𝗳𝘆 — aesthetics, OCR, infographics & editing experts consolidated via Multi-Expert On-Policy Distillation ✨ 𝗥𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 ➡️ 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 — WISE 0.70 ➔ 0.81 with CoT ✨ 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 ➡️ 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 — RealUnify-GEU 47.5 ➔ 56.3 vs. U1-SFT ✨ 𝗩𝗶𝘀𝘂𝗮𝗹 𝗥𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 — 68.2% on VBVR-Pro-Bench, ahead of Nano-Banana-Pro (56.4%) and GPT-Image-2 (50.7%) Full training recipes — SFT, RL & multi-expert on-policy distillation — are open-sourced. 📄 Full Technical Report: https://huggingface.co/papers/2609.11929 🤗 HF: https://huggingface.co/collections/sensenova/sensenova-u15 💻 GitHub: https://github.com/OpenSenseNova/SenseNova-U1 ✏️ SenseNova Studio: https://unify.light-ai.top/ ⚡️Token Plan: https://www.sensenova.ai/token-plan 👾 Discord: https://discord.com/invite/BuTXPHmQub

🌕 This 𝗠𝗶𝗱-𝗔𝘂𝘁𝘂𝗺𝗻 𝗙𝗲𝘀𝘁𝗶𝘃𝗮𝗹, we're sharing a beautiful 𝗔𝗜 𝘃𝗶𝗱𝗲𝗼 𝗺𝗮𝗱𝗲 𝘄𝗶𝘁𝗵 𝗦𝗲𝗻𝘀𝗲𝗧𝗶𝗺𝗲'𝘀 𝘃𝗶𝗱𝗲𝗼 𝗰𝗿𝗲𝗮𝘁𝗶𝗼𝗻 𝗮𝗴𝗲𝗻𝘁, #Seko. It's the work of Chengdu-based AI creator 𝗖𝗵𝗲𝗻 𝗔𝗶𝗷𝘂𝗻. An eyewear shop owner who taught himself AI, Chen has produced more than 200 AI videos, building a following of over 600,000. Here’s to imagination taking flight under the Mid-Autumn moon. ✨

𝗖𝗮𝗽𝘁𝘂𝗿𝗶𝗻𝗴 𝗔𝗜'𝘀 𝗡𝗲𝘅𝘁 𝗪𝗮𝘃𝗲: 𝗙𝗿𝗼𝗺 𝗠𝗼𝗱𝗲𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝘁𝗼 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 At Guotai Junan Securities' 2026 Autumn Conference, Philip Wong, SenseTime's Head of Capital Markets, delivered the opening presentation for the Tech & AI track, outlining three fundamental paradigm shifts reshaping the AI industry: 🔹 𝗙𝗿𝗼𝗺 𝘀𝗶𝗻𝗴𝗹𝗲-𝗺𝗼𝗱𝗮𝗹 𝘁𝗼 𝗻𝗮𝘁𝗶𝘃𝗲 𝗺𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 — perception becomes the foundation of intelligence, as real-world problems demand unified understanding of text, vision and reasoning. 🔹 𝗙𝗿𝗼𝗺 𝘁𝗼𝗸𝗲𝗻 𝗰𝗼𝗻𝘀𝘂𝗺𝗽𝘁𝗶𝗼𝗻 𝘁𝗼 𝘁𝗮𝘀𝗸 𝗱𝗲𝗹𝗶𝘃𝗲𝗿𝘆 — AI is evolving from a tool that answers questions into an agent that completes complex, multi-step workflows end-to-end. 🔹 𝗙𝗿𝗼𝗺 𝘀𝗶𝗻𝗴𝗹𝗲-𝗽𝗼𝗶𝗻𝘁 𝗺𝗼𝗱𝗲𝗹𝘀 𝘁𝗼 𝘀𝘆𝘀𝘁𝗲𝗺-𝗹𝗲𝘃𝗲𝗹 𝗳𝘂𝗹𝗹-𝘀𝘁𝗮𝗰𝗸 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 — deployment and delivery cost, not model benchmarks, define the new competitive moat. SenseTime's "𝗢𝗻𝗲 𝗠𝗼𝗱𝗲𝗹 + 𝗢𝗻𝗲 𝗧𝗼𝗸𝗲𝗻 𝗙𝗮𝗰𝘁𝗼𝗿𝘆 + 𝗢𝗻𝗲 𝗔𝗴𝗲𝗻𝘁 𝗛𝗮𝗿𝗻𝗲𝘀𝘀" 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 is purpose-built for these shifts. We are well positioned to capture the next wave of AI.

Thanks to Flash Cash for reviewing our 𝗼𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗼𝗳𝗳𝗶𝗰𝗲 𝘀𝗸𝗶𝗹𝗹 𝘀𝘂𝗶𝘁𝗲! Instead of switching between separate tools, 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮 𝗦𝗸𝗶𝗹𝗹𝘀 integrates 𝗗𝗲𝗲𝗽 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵, 𝗗𝗮𝘁𝗮 𝗔𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀, 𝗣𝗣𝗧 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻, and 𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻—processing raw data into editable .pptx decks. Built for any skills-compatible agent, including OpenClaw and hermes-agent. 👉 ＧitHub https://github.com/OpenSenseNova/SenseNova-Skills ⚡ Quick installation (bundled with Hermes Agent & OpenClaw) https://github.com/OpenSenseNova/SenseNova-Skills/blob/main/INSTALL.md 👾 Discord https://discord.gg/BuTXPHmQub 🎥 Watch the breakdown: https://m.youtube.com/watch?v=RLvqfyalo-8

“𝗖𝗼-𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝗮𝗻 𝗔𝗜 𝗷𝗼𝘂𝗿𝗻𝗲𝘆 𝘄𝗲 𝘁𝗮𝗸𝗲 𝘁𝗼𝗴𝗲𝘁𝗵𝗲𝗿 𝘄𝗶𝘁𝗵 𝗼𝘂𝗿 𝗽𝗮𝗿𝘁𝗻𝗲𝗿𝘀—𝘀𝗵𝗮𝗿𝗶𝗻𝗴 𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲, 𝗰𝗿𝗲𝗮𝘁𝗶𝗻𝗴 𝗻𝗲𝘄 𝗸𝗻𝗼𝘄-𝗵𝗼𝘄, 𝗮𝗻𝗱 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗶𝗻𝗴 𝘁𝗮𝗹𝗲𝗻𝘁 𝗮𝗹𝗼𝗻𝗴 𝘁𝗵𝗲 𝘄𝗮𝘆.” At the Smart APAC · AI Innovation Summit 2026, hosted by China Telecom in Bangkok, Thailand, Jeff Shi, President of Asia Pacific at SenseTime, shared his perspectives on how the company is creating value for customers and partners in Thailand and across the region. Reflecting on SenseTime’s evolution and the power of collaboration, Jeff noted: “SenseTime started as a Vision AI powerhouse. Today, our vision is to help customers elevate their offerings and capabilities from sensing intelligence to decision intelligence.” 𝗞𝗲𝘆 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀 & 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: 💡 𝗧𝗵𝗲 𝗣𝗼𝘄𝗲𝗿 𝗼𝗳 “𝗖𝗵𝗶𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻”: The global expansion of Chinese technology is driven by two powerful forces: growing international demand for proven applications from China, and an increasingly strong supply of innovative technologies, particularly AI, from Chinese companies. 🚀 𝗦𝗰𝗮𝗹𝗶𝗻𝗴 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗜: Backed by its proprietary AI Cloud and full-stack AI strategy, SenseTime has seen its average daily token output surpass 2.4 trillion, representing a 22-fold increase year over year. 🌏 𝗘𝗺𝗽𝗼𝘄𝗲𝗿𝗶𝗻𝗴 𝗟𝗼𝗰𝗮𝗹 𝗘𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺𝘀: Across Southeast Asia, including Thailand, SenseTime is collaborating with more than 200 enterprise partners to deploy AI-native solutions in smart cities, transportation, banking and financial services, and healthcare—driven by a shared commitment to co-innovation and ecosystem growth. Our sincere thanks to @chinateleglobal and all our industry partners for bringing together a vibrant community of innovators and advancing our shared vision of a more connected, intelligent future for Asia Pacific. #SenseTime #AIInnovation #CoInnovation #AsiaPacific

𝗔 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁𝗶𝗮𝘁𝗲𝗱 𝗰𝗼𝗺𝗺𝗲𝗿𝗰𝗶𝗮𝗹 𝗺𝗼𝗱𝗲𝗹 𝗳𝗼𝗿 𝗔𝗜. As highlighted by @SCMPNews, SenseTime is carving a unique path to profitability — moving beyond the blind chase for model size and focusing on 𝗵𝗲𝗹𝗽𝗶𝗻𝗴 𝗰𝗹𝗶𝗲𝗻𝘁𝘀 𝗰𝗼𝗺𝗽𝗹𝗲𝘁𝗲 𝗿𝗲𝗮𝗹-𝘄𝗼𝗿𝗹𝗱 𝘁𝗮𝘀𝗸𝘀. Our “Models + Token Factory + Agent Harness” strategy is built to continuously 𝗲𝗻𝗵𝗮𝗻𝗰𝗲 𝗔𝗜 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 𝗮𝗻𝗱 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆: 🔹𝗠𝗼𝗱𝗲𝗹𝘀 — advancing the boundaries of intelligence 🔹𝗧𝗼𝗸𝗲𝗻 𝗙𝗮𝗰𝘁𝗼𝗿𝘆 — scaling intelligence production with greater quality and efficiency 🔹𝗔𝗴𝗲𝗻𝘁 𝗛𝗮𝗿𝗻𝗲𝘀𝘀 — connecting models, knowledge, tools and workflows to deliver end-to-end task execution Together, we are 𝘁𝘂𝗿𝗻𝗶𝗻𝗴 𝗔𝗜 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝗶𝗲𝘀 𝗶𝗻𝘁𝗼 𝘀𝘂𝘀𝘁𝗮𝗶𝗻𝗮𝗯𝗹𝗲 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝘃𝗮𝗹𝘂𝗲. Read the full story in SCMP: https://www.scmp.com/tech/tech-trends/article/3366436/how-generative-ai-helps-sensetime-turn-profit-even-chinese-peers-struggle

𝟳𝟰𝗠+ 𝘃𝗶𝗲𝘄𝘀. 𝗢𝗻𝗲 𝗰𝗿𝗲𝗮𝘁𝗼𝗿. 𝗭𝗲𝗿𝗼 𝗽𝗿𝗼𝗳𝗲𝘀𝘀𝗶𝗼𝗻𝗮𝗹 𝗳𝗶𝗹𝗺𝗺𝗮𝗸𝗶𝗻𝗴 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲. 𝗖𝗛𝘂𝗻𝘆𝗲, 𝗮 𝟮𝟮-𝘆𝗲𝗮𝗿-𝗼𝗹𝗱 𝗮𝗿𝘁 𝗱𝗲𝘀𝗶𝗴𝗻 𝘀𝘁𝘂𝗱𝗲𝗻𝘁, 𝘁𝘂𝗿𝗻𝗲𝗱 𝗮 𝗻𝗶𝗴𝗵𝘁𝗺𝗮𝗿𝗲 𝗶𝗻𝘁𝗼 𝗗𝗨𝗛𝗔𝗜, 𝗮 𝗝𝗮𝗽𝗮𝗻𝗲𝘀𝗲-𝘀𝘁𝘆𝗹𝗲 𝘇𝗼𝗺𝗯𝗶𝗲 𝘀𝗵𝗼𝗿𝘁 𝗱𝗿𝗮𝗺𝗮. He doesn't even speak Japanese. He did it with #Seko, SenseTime's AI video creation Agent. 🎬 Watch the trailer and see how Seko helped bring CHUNYE's creative vision to life. From Episode 3 onward, DUHAI has been created entirely with Seko. What used to require a full production crew, CHunye now does on his own. 𝗦𝗲𝗸𝗼 𝘁𝘂𝗿𝗻𝘀 𝗿𝗮𝘄 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗲 𝗶𝗱𝗲𝗮𝘀 𝗶𝗻𝘁𝗼 𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲𝗱 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀, 𝗯𝗿𝗶𝗻𝗴𝗶𝗻𝗴 𝗰𝗵𝗮𝗿𝗮𝗰𝘁𝗲𝗿𝘀, 𝘀𝗰𝗲𝗻𝗲𝘀 𝗮𝗻𝗱 𝗽𝗿𝗼𝗽𝘀 𝘁𝗼𝗴𝗲𝘁𝗵𝗲𝗿 𝗼𝗻 𝗼𝗻𝗲 𝗰𝗮𝗻𝘃𝗮𝘀. 𝗜𝘁𝘀 𝘃𝗶𝘀𝘂𝗮𝗹 𝗺𝗲𝗺𝗼𝗿𝘆 𝗮𝗻𝗱 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 𝗵𝗲𝗹𝗽 𝗸𝗲𝗲𝗽 𝘁𝗵𝗲𝗺 𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝘁 𝗮𝗰𝗿𝗼𝘀𝘀 𝗲𝗽𝗶𝘀𝗼𝗱𝗲𝘀. 𝗕𝘆 𝗘𝗽𝗶𝘀𝗼𝗱𝗲 𝟵, 𝘁𝗵𝗲 𝘀𝗲𝗿𝗶𝗲𝘀 𝗵𝗮𝗱 𝘀𝘂𝗿𝗽𝗮𝘀𝘀𝗲𝗱 𝟳𝟰𝗠 𝗰𝘂𝗺𝘂𝗹𝗮𝘁𝗶𝘃𝗲 𝘃𝗶𝗲𝘄𝘀 𝗼𝗻 𝗗𝗼𝘂𝘆𝗶𝗻 𝗮𝗻𝗱 𝗯𝗲𝘆𝗼𝗻𝗱. That's how AI can turn individual creativity into work that reaches millions — a glimpse of the potential of the One-Person Company (OPC). 𝗜𝗳 𝘆𝗼𝘂 𝗵𝗮𝘃𝗲 𝗮 𝘀𝘁𝗼𝗿𝘆 𝘁𝗼 𝘁𝗲𝗹𝗹, 𝗦𝗲𝗸𝗼 𝗼𝗽𝗲𝗻𝘀 𝘁𝗵𝗲 𝗱𝗼𝗼𝗿 𝘁𝗼 𝘁𝘂𝗿𝗻 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗲 𝗮𝗺𝗯𝗶𝘁𝗶𝗼𝗻𝘀 𝗶𝗻𝘁𝗼 𝗿𝗲𝗮𝗹𝗶𝘁𝘆 — 𝗮𝗻𝗱 𝗼𝗽𝗽𝗼𝗿𝘁𝘂𝗻𝗶𝘁𝘆. Video Credits: CHunye

𝗪𝗶𝘀𝗵𝗶𝗻𝗴 𝘆𝗼𝘂 𝗮𝗻𝗱 𝘆𝗼𝘂𝗿 𝗹𝗼𝘃𝗲𝗱 𝗼𝗻𝗲𝘀 𝗷𝗼𝘆, 𝗽𝗿𝗼𝘀𝗽𝗲𝗿𝗶𝘁𝘆, 𝗮𝗻𝗱 𝗶𝗻𝘀𝗽𝗶𝗿𝗮𝘁𝗶𝗼𝗻 𝘁𝗵𝗶𝘀 𝗠𝗶𝗱-𝗔𝘂𝘁𝘂𝗺𝗻 𝗙𝗲𝘀𝘁𝗶𝘃𝗮𝗹. 🌕

🎬 𝗗𝘆𝗻𝗮𝗺𝗶𝗰 𝗗𝗲𝘀𝗶𝗴𝗻, 𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗯𝘆 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮 𝟲.𝟴 𝗙𝗹𝗮𝘀𝗵 𝗣𝗿𝗲𝘃𝗶𝗲𝘄, 𝗯𝗿𝗶𝗻𝗴𝘀 𝘀𝘁𝗮𝘁𝗶𝗰 𝗶𝗺𝗮𝗴𝗲𝘀 𝘁𝗼 𝗹𝗶𝗳𝗲. Here’s how it works: 🔹𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝘀 how text, subjects, and visual elements relate, and decides what to keep static and what to animate 🔹𝗖𝗵𝗼𝗼𝘀𝗲𝘀 the right form for each element: HTML/CSS, SVG, transparent images, or video 🔹𝗖𝗵𝗼𝗿𝗲𝗼𝗴𝗿𝗮𝗽𝗵𝘀 text reveals, subject motion, and environmental responses into a coherent viewing rhythm 🔹𝗖𝗵𝗲𝗰𝗸𝘀 𝗮𝗻𝗱 𝗿𝗲𝗳𝗶𝗻𝗲𝘀 its output against the source design SenseNova 6.8 Flash is on the way. Stay tuned! Want to try Dynamic Design before launch? We’re offering a 𝗯𝗲𝘁𝗮 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲 with limited access. 𝗟𝗲𝗮𝘃𝗲 𝗮 𝗰𝗼𝗺𝗺𝗲𝗻𝘁 𝗯𝗲𝗹𝗼𝘄 𝘁𝗼 𝗷𝗼𝗶𝗻. 👇

Sometimes, you just have to find your own way out. 🐧 🎬 𝗗𝘆𝗻𝗮𝗺𝗶𝗰 𝗗𝗲𝘀𝗶𝗴𝗻, 𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝗯𝘆 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮 𝟲.𝟴 𝗙𝗹𝗮𝘀𝗵 𝗣𝗿𝗲𝘃𝗶𝗲𝘄, 𝗯𝗿𝗶𝗻𝗴𝘀 𝘀𝘁𝗮𝘁𝗶𝗰 𝗶𝗺𝗮𝗴𝗲𝘀 𝘁𝗼 𝗹𝗶𝗳𝗲.

We release the most comprehensive 𝗲𝗺𝗽𝗶𝗿𝗶𝗰𝗮𝗹 𝘀𝘁𝘂𝗱𝘆 𝗼𝗳 𝗚𝗣𝗧-𝟲-𝗔𝘀𝘁𝗿𝗮’𝘀 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 across 12 simulation benchmarks and 8 real-robot platforms, together with 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮-𝗥𝗼𝗯𝗼𝗥𝗦𝗜: an open-source harness framework and a 27B model that runs at ~1/50 the cost of GPT-6-Astra. Here's what we found: 🔹 𝗧𝗵𝗲 𝗰𝗮𝗽𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗰𝗲𝗶𝗹𝗶𝗻𝗴 𝗶𝘀 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗺𝘂𝗰𝗵 𝗵𝗶𝗴𝗵𝗲𝗿. Under a unified protocol (ICL + up to 10× the interaction budget), GPT-6-Astra achieves clear SOTA across all 12 benchmarks: RoboDojo: 27.6% at 1× → 81.9% at 10×, DexJoCo: 87.3% vs. 40.3% prior SOTA, while the LIBERO benchmarks are pushed to saturation. 🔹 𝗧𝗼𝗼 𝗺𝘂𝗰𝗵 𝗯𝘂𝗰𝗸 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗯𝗮𝗻𝗴. Even a simple task like placing pens in a holder costs ~$55 in inference. 🔹 𝗧𝗵𝗲 𝘄𝗼𝗿𝗹𝗱 𝗱𝗼𝗲𝘀𝗻’𝘁 𝘄𝗮𝗶𝘁. Time matters as real robots overheat, and an egg burned while the model was still thinking. Our takeaway: GPT-6-Astra has largely crossed the threshold of 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗰𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗲, but 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗲𝗺𝗯𝗼𝗱𝗶𝗲𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 is not here yet. 🌐 Project page: https://www.sensenova.ai/has-gpt-6-astra-achieved-embodied-intelligence 📄 Report: https://www.sensenova.ai/gpt-6/assets/docs/has_gpt-6_achieved_embodied_intelligence.pdf 💻 Code & model: https://github.com/OpenSenseNova/SenseNova-RoboRSI

𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮 𝗟𝗼𝗼𝗽-𝟵𝗕-𝗣𝗿𝗲𝘃𝗶𝗲𝘄, 𝗻𝗼𝘄 𝗼𝗽𝗲𝗻 𝘀𝗼𝘂𝗿𝗰𝗲! 𝗦𝗰𝗮𝗹𝗶𝗻𝗴 𝗰𝗼𝗺𝗽𝘂𝘁𝗲 𝗱𝗲𝗽𝘁𝗵, 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝗺𝗼𝗱𝗲𝗹 𝘀𝗶𝘇𝗲: our 9B Looped Transformer uses parameter-shared loops + mHC (Manifold-Constrained Hyper-Connections) to improve information flow and reasoning. 📊 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: 🔹Delivers 𝟮𝟳𝗕-𝗰𝗹𝗮𝘀𝘀 𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 on complex tasks, outperforming Qwen3.5-27B on 𝗪𝗼𝗿𝗸𝗕𝘂𝗱𝗱𝘆 𝗕𝗲𝗻𝗰𝗵, 𝗥𝗮𝗰𝗰𝗼𝗼𝗻𝗣𝗣𝗧 𝗕𝗲𝗻𝗰𝗵 & 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻𝗕𝗲𝗻𝗰𝗵 🔹𝗺𝗛𝗖 𝗯𝗼𝗼𝘀𝘁𝘀 𝘀𝗰𝗼𝗿𝗲𝘀 by 𝟰.𝟰𝟮 𝗽𝗼𝗶𝗻𝘁𝘀 on GPQA Diamond and 𝟭𝟯.𝟯𝟰 on TeamBench 🔹Strong across 𝗰𝗼𝗱𝗶𝗻𝗴, 𝗼𝗳𝗳𝗶𝗰𝗲 & 𝗮𝗴𝗲𝗻𝘁 𝘁𝗮𝘀𝗸𝘀, with 𝗶𝗻𝘁𝗲𝗿𝗮𝗰𝘁𝗶𝘃𝗲, 𝗮𝗻𝗶𝗺𝗮𝘁𝗲𝗱 𝗛𝗧𝗠𝗟 𝗱𝗲𝗺𝗼𝘀 🔬 𝗡𝗲𝘅𝘁: Scaling to larger models, adaptive fast/slow thinking & more efficient reasoning. 🤗 Model: https://huggingface.co/sensenova/SenseNova-6.8-Loop-9B-Preview 💻 GitHub: https://github.com/OpenSenseNova/SenseNova-Loop 👾Discord: https://discord.com/invite/BuTXPHmQub