[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-gpt-5-6-sol-ultra-cdc-conjecture":3,"topics-all":36,"news-related-2c19e48b-438c-4f90-8d60-a1621e4689bf":65},{"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},"2c19e48b-438c-4f90-8d60-a1621e4689bf","GPT-5.6 Sol Ultra 用 64 并发子代理攻下 50 年 Cycle Double Cover 猜想","2026 年 7 月 10 日,OpenAI 研究员 Ethan Knight 在 X 宣布:GPT-5.6 Sol Ultra 用 64 个并行子代理在一小时内给出 Cycle Double Cover 猜想完整证明,同步公开三页证明 PDF 与触发 prompt。这是 CDC 猜想自 1973 年 Szekeres、1979 年 Seymour 独立提出以来,首次公开宣告正面证明。\n\nCDC 问每个无桥图能否找到一组圈,让每条边恰好落入两个圈中。OpenAI 走经典 8-flow 路线:归约到无桥三次图,借既有 nowhere-zero 8-flow 定理拿到边标记,再用新线性代数论证转写为圈分解。整份三页文本几乎只用初等图论与算法式构造,没引新框架——这是这道半世纪悬案第一次以\"短+初等\"形态被击中。\n\n真正可读信号在 prompt 编排:64 子代理被要求\"动态且激进\"协作,前几轮不锁方向,失败代理显式 blocked,对抗型代理专盯病态反例,只有通过自身内部审计的完整证明才允许回传——这把多代理合作直接焊进了研究方法论。\n\nTerminal-Bench 2.1 上 Sol Ultra 拿到 91.9%,比次席 Claude Mythos 5、GPT-5.5 的 88.0% 高出近 4 分,多代理路线在硬结构化问题上的可重复性已经站住脚。下一份 64 子代理输出单的合理候选大概会落在 Erdős–Ginzburg–Ziv、密码学 hardness 上界这类组合结构清晰的悬案上——但数学社群的标准动作仍然是先分组核查三页 PDF,再尝试补一遍 Lean 形式化,在图论领域这条路还要走很长时间。","https:\u002F\u002Fcdn.openai.com\u002Fpdf\u002F04d1d1e4-bc75-476a-97cf-49055cd98d31\u002Fcdc_proof.pdf","15975962-b5fe-49e5-ae68-687ba6cb7015",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"6ad31a14-c0da-42df-81fd-564281f768db","agentic-ai",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",{"id":18,"name":19,"slug":19,"description":13,"color":13},"baf131c1-687a-49f4-87f6-4dd87c1c692f","gpt",{"id":21,"name":22,"slug":22,"description":13,"color":13},"42e59a88-7795-47dc-a334-ef1e72c24347","openai",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"a6465359-43ec-4b73-acc4-1f67a61b383d","en","GPT-5.6 Sol Ultra cracks 50-year cycle double cover conjecture","On July 10, 2026, OpenAI researcher Ethan Knight announced on X: GPT-5.6 Sol Ultra, with 64 parallel sub-agents, delivered a complete proof of the Cycle Double Cover conjecture in one hour, simultaneously publishing a three-page proof PDF and the triggering prompt. This is the first publicly declared positive proof of the CDC conjecture since it was independently proposed by Szekeres in 1973 and Seymour in 1979. The CDC asks whether every bridgeless graph can find a set of cycles such that each edge lies in exactly two cycles. OpenAI takes the classic 8-flow route: reduce to bridgeless cubic graphs, leverage the existing nowhere-zero 8-flow theorem to get edge labels, then use new linear-algebra arguments to rewrite as cycle decomposition. The entire three-page text uses only elementary graph theory and algorithmic construction, with no new framework — this is the first time this half-century-old problem has been hit in \"short + elementary\" form. The real readable signal lies in the prompt orchestration: 64 sub-agents are required to collaborate \"dynamically and aggressively\", the first few rounds don't lock direction, failed agents are explicitly blocked, adversarial-type agents focus on pathological counterexamples, and only complete proofs passing their own internal audit are allowed to return — this welds multi-agent collaboration directly into the research methodology. On Terminal-Bench 2.1, Sol Ultra gets 91.9%, almost 4 points above the second-place Claude Mythos 5 and GPT-5.5's 88.0%, the reproducibility of the multi-agent route on hard structural problems has stood firm. The next reasonable candidate for a 64-sub-agent output sheet will likely fall on combinatorial-structurally clear problems like Erdős–Ginzburg–Ziv or cryptographic hardness upper bounds — but the standard move in the math community is still to first split up to check the three-page PDF, then try to supplement with Lean formalization; in graph theory this road is still long.","gpt-5-6-sol-ultra-cdc-conjecture","2026-07-11T02:01:00Z","2026-07-11T10:13:01.745099Z","2026-08-19T02:08:40.142862Z",true,"agent",582,[37,47,56],{"slug":38,"tag_slug":39,"title_zh":40,"title_en":41,"intro_zh":42,"intro_en":43,"id":44,"is_active":33,"created_at":45,"modified_at":46},"image-generation","text-to-image","图像生成：从文生图到逐框编辑的模型竞赛","Image Generation: from text-to-image to frame-level editing","文生图已是红海，战场正转向编辑与可控性：FLUX.2\u002F3 用开源和量化把生图搬进消费级显卡，Qwen-Image、Seedream 等国产模型紧咬不放，GPT Image 2 则把推理能力带进像素。本专题追踪图像生成与编辑模型的发布、架构演进、加速量化与商业化数据，看「好图」如何变成「可用的图」。","Text-to-image is a crowded race, and the frontier has moved to editing and control: FLUX.2\u002F3 push open weights and quantization onto consumer GPUs, Qwen-Image and Seedream close in from China, and GPT Image 2 brings reasoning into pixels. This topic tracks model releases, architecture shifts, acceleration work, and the business numbers behind \"pretty\" becoming \"usable\".","52de7c59-7864-4569-8ae7-3d656d26a1a1","2026-11-20T07:20:23.534506Z","2026-11-20T07:20:23.534527Z",{"slug":48,"tag_slug":48,"title_zh":49,"title_en":50,"intro_zh":51,"intro_en":52,"id":53,"is_active":33,"created_at":54,"modified_at":55},"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":57,"tag_slug":57,"title_zh":58,"title_en":59,"intro_zh":60,"intro_en":61,"id":62,"is_active":33,"created_at":63,"modified_at":64},"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":66},[67,72,77,82,87,92],{"id":68,"title":69,"news_slug":70,"published_at":71},"9dffcd24-e154-4986-baab-be712e6e0442","GPT-5.6 Sol通关Factorio:低思考档打出火箭","gpt-5-6-sol-factorio-rocket-run","2026-11-01T13:20:38+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"ce0faacb-378d-4478-bd14-99cccc140ea0","GPT-6.1 Terra 首发:十分之一价格追平Astra","gpt-6-1-terra-computer-use","2026-10-26T19:08:07+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"390c2437-4e4f-45ec-8270-67c5bfa4fa47","ChatGPT、Claude、Grok、Gemini 罕见同时下线,周四早晨全球 AI 集体失声","chatgpt-claude-grok-gemini-thursday-outage","2026-09-05T06:00:00+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"9e58d587-3c1b-44c5-ad36-daf23aeb42a2","微软叫停 tokenmaxxing:GitHub Copilot 默认切回 GPT-5.6 Sol,Parikh 设 token 预算","microsoft-token-budget-gpt-5-6-default","2026-09-03T00:30:00+00:00",{"id":88,"title":89,"news_slug":90,"published_at":91},"d1c7b405-fe4e-40f9-9249-a12e2bba6913","GPT-5.6 八月更新：把「推理强度滑块」下放给 Plus\u002FPro，同时把免费用户拉进 Luna 时代","openai-gpt-5-6-august-update-reasoning-slider","2026-08-10T20:00:00+00:00",{"id":93,"title":94,"news_slug":95,"published_at":96},"b9e635eb-ac5e-412d-8904-f113ad3fd5ec","微软宣布工程师 AI token 预算上限并把 OpenAI GPT-5.6 Sol 设为 GitHub Copilot 内部默认模型","microsoft-copilot-gpt-5-6-sol-default-token-budget-0806","2026-08-05T16:00:00+00:00"]