MOSS Image AI
OpenMOSS 的 MOSS Image 是一款开放权重 5B 文生图模型,可清晰呈现中英文文字,并根据 JSON 布局准确放置文字和物体;支持 1K、1.5K 或 2K。 已于 2026年10月 发布。 可在 创作工作台 中在线试用,也可从代码中调用,最高比 the fal API price 低 27%。
- 开发者
- OpenMOSS
- MOSS Image 系列
- 发布时间
- 2026年10月
- Sizes
- up to 2K
- 3 种尺寸
- 起价
- $0.0384 每张图像
- 1K
试用
立即体验 MOSS Image AI
费用从你的余额中扣除,价格显示在按钮上;生成失败会退款。结果会保存到你的账户。
使用 MOSS Image 制作
A hand-painted wooden sign outside a tea shop that reads “山间茶舍 MOUNTAIN TEA HOUSE — OPEN DAILY 9–6”, morning mist, warm light
A café chalkboard menu that reads “今日特饮 TODAY'S SPECIALS — 桂花拿铁 Osmanthus Latte $5”, chalk texture, warm interior light
{"scene_type":"Chinese literature textbook photograph","gen_brief":"A realistic editorial photograph of an open Chinese senior-school literature textbook on a warm wooden study desk. The printed lesson presents Wang Xizhi's Lanting preface in authentic simplified Chinese, with carefully typeset classical prose and compact explanatory footnotes. Use a nearly overhead camera, natural window light, subtle paper translucency, gently curved pages, and soft binding shadows. The image should feel like an ordinary student's clean textbook snapshot rather than a designed poster, with crisp readable printing and restrained, believable photographic texture.","elements":[{"type":"obj","salience":"primary","identity":"Open Chinese literature textbook","shape":"Ivory uncoated paper, gently bowed binding, fine black textbook printing, faint reverse-side ink, layered page edges, understated blue lesson accents.","colors":[{"color":"Ivory paper","hex":"#F5F0E5"},{"color":"Printing black","hex":"#292824"},{"color":"Muted textbook blue","hex":"#496878"}],"bbox":[-685,-544,691,508]},{"type":"text","text":"古代散文阅读","bbox":[-616,-324,-582,-99],"shape":"Small muted-blue Chinese sans-serif characters, restrained tracking and natural square proportions.","is_watermark":false},{"type":"text","text":"兰亭集序","bbox":[-542,-116,-465,281],"shape":"Bold black Song-style Chinese title, prominent triangular serifs, evenly spaced natural-width characters.","is_watermark":false},{"type":"text","text":"王羲之","bbox":[-446,14,-410,152],"shape":"Small centered black Song-style characters with airy spacing.","is_watermark":false},{"type":"text","text":"永和九年,岁在癸丑,暮春之初,会于会稽山阴\n之兰亭,修禊事也。群贤毕至,少长咸集。此地有崇\n山峻岭,茂林修竹,又有清流激湍,映带左右,引以\n为流觞曲水,列坐其次。虽无丝竹管弦之盛,一觞一\n咏,亦足以畅叙幽情。是日也,天朗气清,惠风和畅。\n仰观宇宙之大,俯察品类之盛,所以游目骋怀,足以\n极视听之娱,信可乐也。","bbox":[-357,-319,-90,436],"shape":"Black Song-style body text, regular square glyphs, generous leading, first-line indentation, subtle paper curvature.","is_watermark":false},{"type":"text","text":"夫人之相与,俯仰一世。或取诸怀抱,悟言一室\n之内;或因寄所托,放浪形骸之外。虽趣舍万殊,静\n躁不同,当其欣于所遇,暂得于己,快然自足,不知\n老之将至;及其所之既倦,情随事迁,感慨系之矣。\n向之所欣,俯仰之间,已为陈迹,犹不能不以之兴怀;\n况修短随化,终期于尽!古人云:“死生亦大矣。”\n岂不痛哉!","bbox":[-60,-319,207,436],"shape":"Matching black Song-style body text, first-line indentation, measured line spacing and crisp punctuation.","is_watermark":false},{"type":"text","text":"注释\n〔永和九年〕公元353年。永和,东晋穆帝年号。\n〔会稽山阴〕今浙江绍兴。\n〔修禊〕古代于暮春临水举行的祓除不祥的仪式。\n〔流觞曲水〕把酒杯浮在弯曲的水道中,任其漂流,\n停在谁的面前,谁就取来饮酒。\n〔悟言〕面对面交谈。“悟”同“晤”。","bbox":[330,-319,546,427],"shape":"Compact black Song-style footnotes; slightly bolder heading, hanging entries, smaller natural-width glyphs and even leading.","is_watermark":false},{"type":"text","text":"42","bbox":[615,41,643,82],"shape":"Tiny black serif page number with native numeral proportions.","is_watermark":false},{"type":"obj","salience":"secondary","identity":"Wooden desktop","note":"Provides an everyday study setting.","bbox":[-728,-544,728,544],"shape":"Matte honey-brown wood with restrained grain, minor surface wear and soft natural shadows.","colors":[{"color":"Warm wood","hex":"#B58B61"},{"color":"Dark wood grain","hex":"#896547"}]},{"type":"obj","salience":"secondary","identity":"Narrow facing-page strip and binding gutter","note":"Makes the photographed page clearly part of an open textbook.","bbox":[-673,-544,667,-354],"shape":"Curved ivory page sliver, dark soft binding crease, subtly curled paper and thin layered edges; no readable printing.","colors":[{"color":"Ivory paper","hex":"#F5F0E5"},{"color":"Gutter shadow","hex":"#B6AA97"}]},{"type":"obj","salience":"secondary","identity":"Textbook running-header rule","note":"Restrained printed lesson styling.","bbox":[-567,-321,-563,435],"shape":"Very thin, straight muted-blue printed rule with slight paper-grain absorption.","colors":[{"color":"Muted textbook blue","hex":"#496878"}]},{"type":"obj","salience":"secondary","identity":"Footnote separator","note":"Distinguishes annotations from the classical prose.","bbox":[298,-319,302,-54],"shape":"Short hairline black printing, clean and understated.","colors":[{"color":"Printing black","hex":"#292824"}]}],"must_not":["not a calligraphy scroll or handwritten manuscript","not a promotional poster or infographic","not modern prose substituted for the classical passage"],"palette":["#F5F0E5","#292824","#B58B61","#496878","#B6AA97"],"bbox_coord_space":"abs_centralized"}
这张示例图由 OpenMOSS 使用 MOSS-Image-Base 模型创作,尺寸为 1088 × 1456;在此生成的图像会有所不同。
图像: OpenMOSS · Apache-2.0 · 重新编码为 WebP,并生成 480 px 缩略图
一块破旧的绿色告示牌斜靠在乡间道路旁,牌面密密麻麻印着一篇幽默诙谐的中文短文,标题为“不一定”。
这张示例图由 OpenMOSS 使用 MOSS-Image-Base 模型创作,尺寸为 1184 × 1776;在此生成的图像会有所不同。
图像: OpenMOSS · Apache-2.0 · 重新编码为 WebP,并生成 480 px 缩略图
一张扫描版中文物理练习试卷,页眉写着“物理”,包含题目说明、六道选择题、公式和两幅示意图。
这张示例图由 OpenMOSS 使用 MOSS-Image-Base 模型创作,尺寸为 1056 × 1488;在此生成的图像会有所不同。
图像: OpenMOSS · Apache-2.0 · 重新编码为 WebP,并生成 480 px 缩略图
多层折叠式纸雕立体绘本,放在一张书桌上,背景纯净突出主题,比例 3:2 横版。翻开的跨页呈现「哪吒闹海」的场景:脚踏风火轮的少年、翻涌的层叠海浪与龙宫剪影,所有元素皆可精细折叠组合,呈现逼真细腻的纸张折叠质感与多层景深。构图统一采用正面视角,整体视觉梦幻唯美,色彩缤纷绚丽,充满奇幻生动的故事氛围。
这张示例图由 OpenMOSS 使用 MOSS-Image-Base 模型创作,尺寸为 1536 × 1024;在此生成的图像会有所不同。 点击“试用”会选择 3:2 · 1K,也就是我们提供的最接近尺寸。
图像: OpenMOSS · Apache-2.0 · 重新编码为 WebP,并生成 480 px 缩略图
A cinematic, photorealistic nighttime scene of a small neighborhood convenience store on a quiet, rain-soaked street. The warmly illuminated storefront occupies the right half of the frame, while the street recedes into soft mist on the left. A lone young adult wearing a charcoal coat stands just outside the entrance beneath a transparent umbrella, seen in three-quarter rear view, gazing through the glass. Amber light reveals shelves inside and spills across the sidewalk, mingling with muted teal neon reflections in shallow puddles. Individual raindrops bead on the umbrella and catch the storefront light; fine streaks of rain remain visible against the darker street. A slightly weathered bicycle leans beside the entrance. Distant apartment windows and a few out-of-focus streetlights add depth without distracting from the figure. Eye-level wide shot, 35mm lens aesthetic, natural perspective, balanced negative space, restrained teal-and-amber color grading, realistic wet asphalt and glass textures, soft highlight bloom, subtle film grain, and gentle atmospheric depth. Quiet, intimate, and nostalgic rather than dramatic; believable lighting, with the figure clearly silhouetted against the warm window. No prominent readable lettering, logos, or watermarks.
这张示例图由 OpenMOSS 使用 MOSS-Image-Base 模型创作,尺寸为 1680 × 944;在此生成的图像会有所不同。 点击“试用”会选择 16:9 · 1K,也就是我们提供的最接近尺寸。
图像: OpenMOSS · Apache-2.0 · 重新编码为 WebP,并生成 480 px 缩略图
提示词示例
A wide view of a refrigerator beside a kitchen counter with a fruit bowl. A small sticky note on the refrigerator door is handwritten in Chinese: “记得买牛奶和鸡蛋!牛奶两盒,鸡蛋一打。”
Minimalist travel poster, cream background, a flat blue mountain illustration, the title “EXPLORE” in large dark-blue capitals at the top
简介
什么是 MOSS Image?
双语文字,清晰准确
MOSS Image 专为文字排版训练:可在招牌、海报、菜单、便笺和包装上准确呈现中英文、换行和混合文字。请将要渲染的文字放在引号中。
以布局作为输入
除了普通提示词和标签列表,MOSS Image 还支持 JSON 布局:包含场景说明和若干项目,每个项目都有描述、要渲染的文字,以及以画布中心为原点的像素边界框。请使用与你请求的分辨率和宽高比相对应的画布;各选项会列出可用画布。
开放权重
OpenMOSS 以 Apache-2.0 许可发布 MOSS Image:这是一款 5B 扩散 Transformer,使用 Qwen3.5 文本编码器,可接受较长提示词。该模型通过面向开源模型的 GPU 推理运行。
功能
MOSS Image 能做什么
双语文字
中英文
JSON 布局
通过边界框指定文字和物体
1K 至 2K
15 种宽高比,每种价格相同
开放权重
Apache-2.0
使用场景
人们用 MOSS Image 做什么
- 根据描述生成图片
- 插画、产品视觉素材和概念图,最高可达 2K。
- 适配任意画幅
- 支持 15 种画幅比例,从横幅到竖屏帖子。
对比情况
MOSS Image 与相似模型对比
MOSS Image 与 Qwen-Image 3.0
MOSS Image
- 宽高比选择更多:15,另一款为 7。
Qwen-Image 3.0
- 价格更低:$0.0228 每张图像,而 1K 时为 $0.0384 每张图像(便宜 41%)。
- 支持编辑,最多可输入 3 张图像。
- 支持 图像编辑。
使用方式
使用 MOSS Image 的三种方式
常见问题
MOSS Image AI 常见问题
MOSS Image 是 OpenMOSS 的开源权重模型。CheaperGen 通过面向开源模型的 GPU 推理提供其开源权重的 API 访问,但不拥有该模型。商标归各自所有者所有。
