[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-maliang-harness-p2v-gap":3,"topics-all":38,"news-related-5757433d-e8ca-4371-89bc-f563dacc8089":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},"5757433d-e8ca-4371-89bc-f563dacc8089","程序能跑≠画对了:MaLiang框架专治AI作画跑偏","9 月 28 日 arXiv 论文 MaLiang-Harness 定义「程序-视觉」鸿沟:模型写的画图代码能跑通,构图、外观、运动却可能全错。框架以生成、检查、修订闭环推进,GPT-6-Astra 生成成功率 100%,96.0% 图像任务全指标达标,视频仅 76.9%;通用能力分预测不了视觉生成表现。","让模型写代码画图,有一个常被忽略的坑:代码能跑,画却是错的。构图偏了、颜色不对、运动轨迹完全跑偏——程序在语法层面无懈可击,视觉层面一塌糊涂。9 月 28 日挂上 arXiv 的 MaLiang-Harness 论文给这个现象起了名字:Program-to-Visual(P2V)鸿沟,并顺手搭了一整套生成-质检-修订流水线,登上 Hugging Face 日榜第二名。\n\n## 什么是 P2V 鸿沟\n\n可执行程序能显式控制图像和视频的构造方式,这是它比提示词强的地方。但生成可运行的代码只是视觉创作的起点:程序完全可能在无报错运行的同时,违反要求的构图、外观或运动。论文把「代码正确」与「画面达标」之间的落差定义为 P2V 鸿沟。这个定义本身就有价值——此前大家笼统说「模型画得不好」,现在有了可度量的切面:生成成功、视觉质量、计算成本三个维度分开算。\n\n## 解法:生成-检查-修订闭环\n\nMaLiang-Harness 让多模态大模型编写绘图与动画程序,经 Canvas、SVG、Three.js 等后端渲染成 PNG 图像或无声 MP4 视频,再检查输出、修订代码,循环推进。三个机制撑起闭环:PEG 持久化保存程序与任务上下文,每次提交带修订快照;TGP 把每次编辑与渲染证据关联留痕;REV 支持从历史修订恢复重跑,并在导出前复核当前修订。项目基于 Deep Agents 构建,带 CLI、聊天界面与批量评测工具,代码开源在 GitHub。\n\n## 实测:100% 能生成,76.9% 才达标\n\n团队在自建的 MaLiang-IBench(50 个图像任务)上评了 11 个闭源 MLLM,在 MaLiang-VBench(13 个视频任务)上评了 4 个。GPT-6-Astra 两个基准都拿到 100% 生成成功,96.0% 的图像任务与 76.9% 的视频任务通过全部质量阈值。更扎心的是另一个发现:通用能力分数与视觉生成表现明显错位,总分相近的模型,满足视觉需求的能力差距悬殊——通用榜单在这里几乎失去预测力。\n\n## 所以呢\n\n两条判断。其一,可编程生成是扩散模型之外的另一条路线:要精确控制排版、结构、运动,写代码天然比采样可控,前提是补上「自查-修订」这个环节,MaLiang 把它工程化了。其二,选型别再迷信通用跑分,视觉生成这种复合能力,得用任务专属基准说话。项目取名「马良」,中文语境很难不想到神笔马良——神笔的从来不是笔,是那双会看、会改的眼睛。\n\n参考:[arXiv 论文](https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.34309)、[GitHub 仓库](https:\u002F\u002Fgithub.com\u002Fgulucaptain\u002FMaLiang-Harness)。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.34309","7437aeb9-930c-4866-a2e9-48003c1a792b",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"120fa59a-ff6f-4537-9bf5-f818df636a0e","benchmark",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":19,"name":20,"slug":20,"description":14,"color":14},"c883fd20-1d66-4fb7-9fc7-320fa7f87023","text-to-image",{"id":22,"name":23,"slug":23,"description":14,"color":14},"ebe5dcd1-46b1-4298-b8c2-8e0e2f456e56","video-generation",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"c730f7b3-3874-43d2-bb1f-7f588e1c26e1","en","Code Runs, Image Wrong: MaLiang Tames the P2V Gap","arXiv paper MaLiang-Harness names the Program-to-Visual gap — code runs, image wrong — and closes it with a draft-inspect-revise loop for visual programs.","Letting models write code to draw has a pitfall everyone underestimates: the code runs, but the picture is wrong. Composition drifts, colors misfire, motion goes off track — the program is flawless at the syntax level and a mess at the visual level. A paper posted to arXiv on September 28 gives the phenomenon a name — the Program-to-Visual (P2V) gap — and ships a full construction-inspection-revision pipeline alongside it, reaching #2 on Hugging Face's daily paper ranking.\n\n## What the P2V gap is\n\nExecutable programs offer explicit control over how images and videos are constructed — that is their edge over prompts. But generating runnable code is only the beginning of visual creation: a program can execute without errors while violating the requested composition, appearance, or motion. The paper defines the discrepancy between \"code correct\" and \"picture acceptable\" as the P2V gap. The definition itself is useful: instead of vaguely saying a model \"draws poorly\", you get measurable slices — generation success, visual quality, and computational cost, tracked separately.\n\n## The fix: a draft-inspect-revise loop\n\nMaLiang-Harness has multimodal LLMs write drawing and animation programs, renders them to PNG images or silent MP4 videos through backends such as Canvas, SVG, and Three.js, then inspects the output and revises the code in a loop. Three mechanisms hold it together: PEG persists programs and task context, snapshotting every committed revision; TGP ties each edit to its rendered evidence; REV restores from historical revisions and re-checks the current one before export. Built on Deep Agents, the project ships a CLI, a chat interface, and batch evaluation tools, with code open-sourced on GitHub.\n\n## Results: 100% can generate, 76.9% actually pass\n\nThe team evaluated 11 closed-source MLLMs on their MaLiang-IBench (50 image tasks) and 4 models on MaLiang-VBench (13 video tasks). GPT-6-Astra hits 100% generation success on both benchmarks, with 96.0% of image tasks and 76.9% of video tasks passing all quality thresholds. The stinging part is the other finding: general capability scores mismatch visual-generation performance — models with similar overall scores differ substantially in satisfying visual requirements. Generic leaderboards lose nearly all predictive power here.\n\n## So what\n\nTwo takeaways. First, programmable generation is a real alternative to diffusion: when you need precise control over layout, structure, and motion, writing code is inherently more controllable than sampling — provided you close the self-check-and-revise loop, which MaLiang engineers into place. Second, stop trusting general benchmarks for model selection; composite abilities like visual generation demand task-specific benches. The project is named after Ma Liang — in Chinese context, the boy with the magic brush. The magic was never the brush; it was the eye that looks, and revises.\n\nReferences: [arXiv paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.34309), [GitHub repo](https:\u002F\u002Fgithub.com\u002Fgulucaptain\u002FMaLiang-Harness).","maliang-harness-p2v-gap","2026-09-30T15:20:00Z","2026-09-30T15:17:19.938024Z","2026-09-30T15:17:19.938041Z",true,"agent",452,[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},"240d8a4c-c025-4080-be14-06099f650a71","xAI 把视频价格打到底:1080p 蒸馏版 Lite 比 Veo 3.1 便宜 65%","grok-imagine-1-5-lite-cheaper-than-veo","2026-10-10T08:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"ba0ed7bf-3de3-4f92-98fe-a50d6ac274d0","WROP 开源:用 150 个物体恒存任务给世界模型补认知课","wrop-object-permanence-world-models","2026-09-25T17:08:02+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"aa9b279e-07c6-4d0c-86f0-403c4af321fa","蚂蚁 RULER:六维评分表给 SVG 生成重写奖励信号","ruler-rubric-rewards-svg-generation","2026-09-23T23:06:24+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"371dd90f-e8bb-46a1-967a-346c28b393df","MiniMax-H3 会不会推理物理世界?517 题实测只对 42%","minimax-h3-physical-reasoning-eval","2026-09-19T17:11:04+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"b95b93e8-294a-4c5b-b53d-ce6ea07c1519","SemComp-Bench 登顶 Hugging Face 日榜:视频生成开始考「任务做没做成」","semcomp-bench-video-task-completion","2026-08-20T13:30:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"e795ec57-7401-458d-a67f-cd18098b2cf3","OpenCoF 把视频生成变成\"显式推理机\":字节 + 港中文用 17K 数据让 Wan 学会\"链帧思考\"","opencof-wan-video-reasoning","2026-07-11T18:01:00+00:00"]