[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-ant-ling-3-1-flash-560b":3,"topics-all":38,"news-related-2a44e5b9-1d17-493e-b0a6-feca88946ec9":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"2a44e5b9-1d17-493e-b0a6-feca88946ec9","蚂蚁 Ling-3.1-flash:效率线拉到 560B","蚂蚁 inclusionAI 发布 Ling-3.1-flash:560B 总参 MoE,激活 25B,较上代翻四倍。定位 agent 与办公,目标 1M 上下文,试用先限 256K,经 Vercel AI Gateway 免费两周,官方称试用后开源。自报五项 agent 基准聚合 81.0。","9 月 29 日,蚂蚁集团 inclusionAI 上线 Ling-3.0-flash 的继任者 Ling-3.1-flash。这款效率定位 MoE 总参拉到约 560B,每 token 激活约 25B——对比上代 124B 总参 \u002F 5.1B 激活,总参翻四倍多,激活接近五倍。官方同步承诺:试用期满后开源权重。\n\n## 从 124B 到 560B,效率线也在堆参数\n\nLing-3.0-flash 当年的卖点是\"小激活跑生产级 Agent\":124B 总参只激活 5.1B,部署成本低。Ling-3.1-flash 把两个数字一起放大——总参 560B、激活 25B——稀疏比例仍约 1\u002F22,但成本档位已完全不同。信号很直白:纯靠小激活参数,agent 任务的质量天花板顶不住了。蚂蚁将其定位在 agent 任务、搜索、办公和专业应用,带显式 thinking mode,瞄准长程任务。\n\n上下文方面,官方目标 1M token,但当前经 Vercel AI Gateway(Novita 托管)的两周免费试用只开 256K 上下文,输出上限 32,768 token(TechNode 报道见 https:\u002F\u002Ftechnode.com\u002F2026\u002F09\u002F30\u002Fant-group-launches-ling-3-1-flash-with-560-billion-parameters\u002F )。\n\n## 先试用、后开源的发布节奏\n\n这次发布最有意思的不是参数,而是节奏:模型先以两周免费试用上线 Vercel AI Gateway,权重、许可证、定价在发布时全部缺席,inclusionAI 只承诺试用结束后开源。第三方追踪站 LLM Releases 在模型页明确标注\"weights not yet posted\"( https:\u002F\u002Fwww.llm-releases.com\u002Fmodels\u002Fling-3-1-flash ),OrcaRouter 的评论更直接:这是一个\"announced model\"——宣布了,但还摸不着。\n\n对照上代直接 MIT 开源上架 Hugging Face,这代先攒真实负载与反馈、再放权重的路径更像商业模型的发布节奏——可以解读为打磨质量,也可以看作开源策略的试探。\n\n## 跑分要打折扣看\n\n官方公布的基准:五项 agent 基准聚合 81.0、Terminal-Bench 4 40.4%、SWE-Atlas 55.9%、HealthBench Professional 65.3%;官方账号另给出 GDPVal-AA v2.1 1,673 Elo、FrontierSWE 75.16(官方公告帖见 https:\u002F\u002Fx.com\u002FAntLingAGI\u002Fstatus\u002F2105335205741596911 )。作为参照,Claude Sonnet 5.5 在同一 Terminal-Bench 4.0 上是 70.6%,Ling-3.1-flash 的 40.4% 距第一梯队还有明显差距;HealthBench Professional 65.3% 倒切中蚂蚁自己的医疗与金融场景。所有数字目前只有蚂蚁自己跑过,无独立复测。\n\n## 所以呢\n\nLing-3.1-flash 值不值得盯,取决于两周后权重是否如约落地、以什么许可证落地。560B 总参的自托管门槛已到多卡集群级别,\"开源\"对多数开发者会从\"我能跑\"变成\"我能微调、蒸馏\"。若真放 MIT,等于送社区一个大号底座;若附带商用条款,这次试用就是纯营销。答案两周内见分晓。","https:\u002F\u002Fx.com\u002FAntLingAGI\u002Fstatus\u002F2105335205741596911","cbed215e-2b0b-4d0d-abf6-13d3df00da9a",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"471c51be-e620-49df-bd6c-0b5504f53f00","ant-group",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",{"id":19,"name":20,"slug":20,"description":14,"color":14},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":22,"name":23,"slug":23,"description":14,"color":14},"d11f0044-8aef-487c-bebe-89ce4683a4a3","moe",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"ceeaef16-4794-4104-9ca3-5bc68c8299b4","en","Ant's Ling-3.1-flash: 560B MoE, try-then-open","Ant's inclusionAI launches Ling-3.1-flash: a 560B MoE with ~25B active per token, 4x its predecessor. Free trial at 256K context, open weights after.","On September 29, Ant Group's inclusionAI lab shipped Ling-3.1-flash, successor to Ling-3.0-flash. The efficiency-focused MoE model scales to roughly 560B total parameters with about 25B activated per token — versus the previous generation's 124B total \u002F 5.1B active, total parameters have grown more than fourfold and active parameters nearly fivefold. The official promise that came with it: open weights after the trial period.\n\n## From 124B to 560B: even the efficiency line is stacking parameters\n\nLing-3.0-flash made its name on \"small activation for production agents\": 124B total parameters with only 5.1B active kept deployment costs low. Ling-3.1-flash blows up both numbers — 560B total, 25B active — keeping the sparsity ratio around 1\u002F22 but landing in a completely different cost class. The industry signal is blunt: pure small-activation designs have hit a quality ceiling on agent tasks. Ant positions the model for agent tasks, search, office software, and specialist applications, with an explicit thinking-mode switch aimed at long-horizon work.\n\nOn context, the stated target is up to 1M tokens, but the current two-week free trial via Vercel's AI Gateway (hosted by Novita) opens only 256K, with output capped at 32,768 tokens (TechNode: https:\u002F\u002Ftechnode.com\u002F2026\u002F09\u002F30\u002Fant-group-launches-ling-3-1-flash-with-560-billion-parameters\u002F ).\n\n## A try-first, open-later release rhythm\n\nThe most interesting thing about this launch isn't the parameters — it's the sequence. The model arrived as a two-week free trial on Vercel's AI Gateway; weights, license, and official pricing were all absent at announcement, with inclusionAI only committing to open-source after the trial ends. The independent tracker LLM Releases explicitly marks \"weights not yet posted\" on the model page ( https:\u002F\u002Fwww.llm-releases.com\u002Fmodels\u002Fling-3-1-flash ), and OrcaRouter's commentary is blunter: this is an \"announced model\" — declared, but not yet touchable.\n\nCompare that with Ling-3.0-flash, which went straight to Hugging Face under MIT. Gathering real-world load and feedback before releasing weights looks more like a commercial-model playbook. You can read it as quality polishing — or as a probe of open-source strategy.\n\n## Read the benchmarks with a discount\n\nOfficially reported scores: an 81.0 aggregate across five agent benchmarks, Terminal-Bench 4 at 40.4%, SWE-Atlas 55.9%, HealthBench Professional 65.3%; Ant Ling's account on X adds GDPVal-AA v2.1 at 1,673 Elo and FrontierSWE at 75.16 (official announcement: https:\u002F\u002Fx.com\u002FAntLingAGI\u002Fstatus\u002F2105335205741596911 ). For reference, Claude Sonnet 5.5 scores 70.6% on the same Terminal-Bench 4.0 — Ling-3.1-flash's 40.4% sits well behind the frontier tier, while HealthBench Professional 65.3% maps neatly onto Ant's own healthcare and finance verticals. Every number so far has been run by Ant alone, with no independent replication.\n\n## So what\n\nWhether Ling-3.1-flash is worth watching comes down to what lands after two weeks: real weights, and under what license. A 560B open model puts self-hosting into multi-GPU-cluster territory, so \"open source\" for most developers degrades from \"I can run it\" to \"I can fine-tune or distill it.\" If it truly ships under MIT, the community gets a large base model for free; if it arrives with commercial strings, this trial was pure marketing. The answer arrives within two weeks.","ant-ling-3-1-flash-560b","2026-10-01T23:15:00Z","2026-10-01T23:17:33.482908Z","2026-10-01T23:17:33.482918Z",true,"agent",863,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"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":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"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":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"b44fb160-9027-44cd-b355-a66869004abc","Nemotron 3.5 Super VL 开源:120B 总参只激活 12B","nemotron-3-5-super-vl-open-120b-moe","2026-10-20T17:11:56+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"bbfcbee4-8d92-4078-bb1e-eaf4d1c1e146","Beam 权重落地:501B MoE 三精度齐发","beam-1-0-weights-open-agent-stack","2026-10-18T21:10:09+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"d529114c-bfd7-4efb-9af4-444d3065b397","Step 5 Preview 开源,2.7 美元\u002F百万输出","stepfun-step-5-preview-600b-moe","2026-10-17T21:12:04+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"cee60a45-f212-4a78-83dd-0226c07b3460","腾讯开源记忆写入模型 Hy-Memory:30B MoE 只激活 3B,四框架平均分超 Opus 5","tencent-hy-memory-30b-memory-writer","2026-10-17T15:14:21+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"ac666c4e-4756-43d8-840a-4eacbf88f804","Meta 开源 816B MoE 模型 Muse Spark Open,跑在两台 Mac Studio 上","meta-muse-spark-open-816b-mac-studio","2026-10-17T03:00:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"1b6ec126-cba1-47da-870b-da19cc523763","Meta 开源 816B MoE 模型 Muse Spark Open:1M 上下文,两台 Mac Studio 跑得动","meta-muse-spark-open-816b-moe-1m-mac-studio","2026-10-15T11:00:00+00:00"]