最新资讯
bloomberg.com 台积电据报考虑生产先进 AI 存储芯片 www.infoq.com Yelp 为 AI Assistant 融合原生与服务器驱动 UI computerworld.com 英国监管机构:企业须为自家 AI 代理的错误负责 siliconangle.com Atomicwork 推出面向企业 IT 与安全的认证 AI 同事市场 computerworld.com OpenAI 收购小型软件公司以改进 Codex bdtechtalks.com Next-Concept Modeling 如何重新思考 AI 文本生成 computerworld.com Box 发布 Architect:用自然语言构建业务流程工作流 techrepublic.com Perplexity 推出 Teammates:每月 200 美元的 AI Agent 如何工作 the-decoder.com METR 称 Anthropic 或已在 7 月越过自身 AI 加速预警阈值 towardsdatascience.com 用 Ollama 和 Python 构建你的第一个多工具 AI Agent the-decoder.com Thinking Machines 的 Inkling 2 Small 称霸西方开放权重模型 dev.to 为 CI/CD 构建 AI 质量门禁
多元动态
登录 注册
大模型 siliconangle.com · @Maria Deutscher

Moonshot AI 发布 3T 参数新旗舰模型 Kim K3.1 Pro

原文 我的收藏

Moonshot AI 推出 Kim K3.1 Pro,这是一款拥有 3 万亿参数的大语言模型,为今年 7 月发布的 Kimi K3 的后训练升级版本。报道称该模型在部分任务上表现优于 Claude Opus 5.5 和 GPT-6 Astra,并提到前代 Kimi K3 由 896 个独立神经网络组成。具体基准与测试条件在摘要中未披露。

AI推荐理由 这是科技媒体对 Moonshot AI 新旗舰模型的报道,给出了具体参数规模(3T)和与前代 Kimi K3 的关系(后训练升级版),并称部分任务表现优于 Claude Opus 5.5 和 GPT-6 Astra。新模型发布属于模型层重要动态;但摘要中的性能对比来自媒体转述,未给出基准名称、测试条件或官方数据,需以官方发布为准。

72

正文

UPDATED 20:55 EDT / OCTOBER 23 2026 AI Moonshot AI launches new flagship Kim K3.1 Pro model with 3T parameters by Maria Deutscher Moonshot AI today introduced Kim K3.1 Pro, a large language model that can perform some tasks better than Claude Opus 5.5 and GPT-6 Astra. The LLM is a post-trained, or upgraded, version of an algorithm that the Chinese artificial intelligence company released in July. The latter model, Kimi K3, comprises 896 individual neural networks with 3 trillion parameters. It only activates 16 of those neural networks at any given time to minimize inference costs. Moonshot AI also optimizes its models’ hardware footprint in other ways. Kim K3 features Kimi Delta Attention, a custom implementation of a memory-saving technology called Gated DeltaNet. It reduces the amount of RAM that the model’s attention mechanism uses to interpret user prompts. The feature was presumably carried over to Kim K3.1 Pro. The company says the new model’s post-training workflow consumed significantly more processing power than similar projects it completed in the past. Additionally, it ran on an upgraded version of the company’s post-training stack. Moonshot AI added automated quality checks and boosted the underlying infrastructure’s efficiency. One area where Kimi K3.1 Pro delivers particularly competitive output quality is 3D modeling. There are two ways to create a 3D model of an object: engineers can draw it or they can write a code snippet that specifies its geometry in numerical terms. Kimi K3.1 Pro takes the latter approach. The LLM scored 92.7% on a 3D model coding benchmark called BenchCAD, which put it within four percentage points of Astra and Claude Opus 5.5. Kimi K3.1 Pro can not only generate static 3D models but also animate them. In one test, Moonshot AI used the model to generate an animated simulation of human motion. Another internal evaluation saw the company create multiple 3D video games. Games are difficult for AI models to generate because they contain a large amount of code and even a single bug can crash them. Pro scored 65.7% on the industry standard Terminal-Bench 4.0 coding benchmark, which put it more than 7% ahead of Astra but slightly behind Opus 5.5. Furthermore, it managed to outperform both models on OfficeQA Pro v2. It’s a test that measures LLMs’ ability to perform complex financial analysis tasks. The model’s performance on popular cybersecurity benchmarks significantly trails the scores of Astra and Opus 5.5. According to Moonshot AI, its engineers deliberately removed cybersecurity-related information from the model’s post-training dataset to address safety risks. The company plans to equip its future models with more advanced vulnerability analysis features. Kimi K3.1 Pro is available via Moonshot AI’s paid cloud services on launch. The company plans to open-source the model’s weights by Nov. 4. Image: Unsplash A message from John Furrier, co-founder of SiliconANGLE: Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities. 15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more 11.4k+ theCUBE alumni — Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network Are you an AWS customer?  Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/ About SiliconANGLE Media SiliconANGLE Media is a recognized leader in digital media innovation, uniting breakthrough technology, strategic insights and real-time audience engagement. As the parent company of SiliconANGLE, theCUBE Network, theCUBE Research, CUBE365, theCUBE AI and theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI. Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations. LATEST STORIES Anthropic reveals Mythos 5 tried submitting malicious code to an open-source project Three insights you may have missed from theCUBE’s coverage of Oracle’s “AI Cyberattacks Are Escalating: How to Secure Your Data Now” virtual event AI drug development startup Pathos AI closes $237M investment Hybrid infrastructure decisions put AI costs and control in focus What to expect during Intuit Connect: Join theCUBE Oct. 28 What to expect during Oracle AI World: Join theCUBE Oct. 28 LATEST STORIES Anthropic reveals Mythos 5 tried submitting malicious code to an open-source project SECURITY - BY MARIA DEUTSCHER . 2 HOURS AGO Three insights you may have missed from theCUBE’s coverage of Oracle’s “AI Cyberattacks Are Escalating: How to Secure Your Data Now” virtual event SECURITY - BY VICTORIA GAYTON . 2 HOURS AGO AI drug development startup Pathos AI closes $237M investment AI - BY MARIA DEUTSCHER . 5 HOURS AGO Hybrid infrastructure decisions put AI costs and control in focus AI - BY JONATHAN ANTHONY . 5 HOURS AGO What to expect during Intuit Connect: Join theCUBE Oct. 28 AI - BY DEVONY HOF . 5 HOURS AGO What to expect during Oracle AI World: Join theCUBE Oct. 28 AI - BY MARK ALBERTSON . 6 HOURS AGO

来源报道 · 1 篇 · 以下链接均前往原始媒体

← 返回精选动态 原文

更多内容