[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-ant-ring-zero-trillion-zero-rl":3,"topics-all":36,"news-related-1805bc3b-3c32-4807-a07e-2b0ab5105015":55},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":13,"view_count":35},"1805bc3b-3c32-4807-a07e-2b0ab5105015","蚂蚁百灵 Ring-Zero 把零强化学习推到万亿级：1T 参数自发涌现出「自我验证」「平行推理」","蚂蚁百灵 InclusionAI 联合 Wayne Xin Zhao 团队 7 月 14 日发布 Ring-Zero 论文,把 zero RL(零强化学习)训练首次推到 1T 参数规模,产物 Ring-2.5-1T-Zero 在七项数学基准上取得有竞争力表现,checkpoint 同步开源到 Hugging Face。zero RL 不依赖人类标注,只用可验证奖励在冷启动模型上直接激发链式思考。过往该路线受算力约束,实验多停在 7B 到 32B 区间;Ring-Zero 通过 clipped importance sampling、训练-推理比值校正、混合精度控制等系统工程把规模推到 1T。论文最有价值的部分不是「能做」,而是「做出来什么样」。作者给出三点:1T 规模显著提升样本效率与性能上限;训练分「发现」与「锐化」两阶段;模型自发涌现拟人化表达、结构化排版、自验证、平行推理、context anxiety 等高级认知行为,让手工设计的奖励启发式变得冗余。针对 CoT 质量,作者提出「可读性、可复现性、效率」三维框架,比单纯比对最终答案更贴近工业部署。这是继 DeepSeek R1 之后「后训练即产品力」叙事的又一次大型验证,1T 级 zero RL 的可行性被打开,后续 MoE、长上下文、推理成本控制等组合拳值得继续观察。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.12395","7437aeb9-930c-4866-a2e9-48003c1a792b",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"471c51be-e620-49df-bd6c-0b5504f53f00","ant-group",{"id":18,"name":19,"slug":19,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",{"id":21,"name":22,"slug":22,"description":13,"color":13},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"ea79e755-a453-4b08-8391-515f69b227db","en","Ant Ring-Zero: zero-RL at 1T params, self-verification emerges","Ant Group's Bailing InclusionAI, together with Wayne Xin Zhao's team, released the Ring-Zero paper on July 14, pushing zero-RL training to the 1T-parameter scale for the first time; the resulting Ring-2.5-1T-Zero achieves competitive performance on seven math benchmarks, and checkpoints are open-sourced to Hugging Face. Zero-RL doesn't rely on human annotation, but uses verifiable rewards to directly stimulate chain-of-thought on a cold-start model. Previously constrained by compute, this path had mostly stopped at the 7B–32B range; Ring-Zero pushes the scale to 1T through clipped importance sampling, training-inference ratio correction, mixed-precision control, and other systems engineering. The most valuable part of the paper is not \"it can be done\", but \"what the result looks like\". The authors give three points: the 1T scale significantly improves sample efficiency and the performance ceiling; training is split into a \"discovery\" phase and a \"sharpening\" phase; the model spontaneously emerges human-like expression, structured layout, self-verification, parallel reasoning, context anxiety, and other high-level cognitive behaviors, making hand-designed reward heuristics redundant. Regarding CoT quality, the authors propose a \"readability, reproducibility, efficiency\" three-dimensional framework, which is closer to industrial deployment than just comparing final answers. This is another large-scale validation of the \"post-training is product power\" narrative after DeepSeek R1; the feasibility of 1T-level zero-RL is opened up, and subsequent combinations like MoE, long context, and reasoning cost control are worth continuing to watch.","ant-ring-zero-trillion-zero-rl","2026-07-15T06:20:00Z","2026-07-15T06:22:15.953674Z","2026-08-19T02:08:40.142862Z",true,"agent",284,[37,46],{"slug":38,"tag_slug":38,"title_zh":39,"title_en":40,"intro_zh":41,"intro_en":42,"id":43,"is_active":33,"created_at":44,"modified_at":45},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":47,"tag_slug":47,"title_zh":48,"title_en":49,"intro_zh":50,"intro_en":51,"id":52,"is_active":33,"created_at":53,"modified_at":54},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":56},[57,62,67,72,77,82],{"id":58,"title":59,"news_slug":60,"published_at":61},"8def771a-d936-4859-930d-02c3011dc55c","LimiX-2 开源：一个模型吃下分类回归插补，表格三榜登顶","limix-2-tabular-foundation-model","2026-09-17T21:09:27+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"176b4807-da61-479f-a514-9381cd13319e","SP3O:3 个锚点修复 PPO critic 的平坦化","sp3o-sparse-critic-supervision","2026-09-17T17:10:01+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"7bae3d71-a5c2-4588-95e7-b5d4b5c7085a","开源模型 4.4 个月追上闭源前沿:Hugging Face 被 NVIDIA 129 亿美元收编","nvidia-acquires-hugging-face-open-source-ai","2026-09-17T08:00:00+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"d41175a7-ad10-4e00-9017-a148fa0a77b3","BenchMIRT 把 LLM 基准拆到单题:Ai2 想让模型排名不再「一张考卷定生死」","ai2-benchmirt-llm-benchmark-audit","2026-09-10T11:05:05+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"f2f0017f-449d-492b-b9d7-90a2013023fb","英伟达 129 亿美元收购 Hugging Face 接近敲定:开源 AI 仓库终被算力霸主收编","nvidia-12-9-billion-hugging-face-acquisition","2026-09-04T03:00:00+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"932ae5e5-3552-4f9d-a6fe-26eedca0bb2b","蚂蚁首个金融增强模型开源在即:Ling-3.0-flash-Fin 押注投研 Agent","ant-ling-3-flash-fin-finance-llm","2026-08-31T19:15:00+00:00"]