Skip to content

Commit 5a4186f

Browse files
gujieyegujieye
committed
chore: update bailian-docs-llm-wiki (2026-08-05)
1 parent 319748e commit 5a4186f

132 files changed

Lines changed: 5173 additions & 16440 deletions

File tree

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

‎skills/bailian-docs-llm-wiki/llms.txt‎

Lines changed: 606 additions & 619 deletions
Large diffs are not rendered by default.

‎skills/bailian-docs-llm-wiki/models/families.jsonl‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -70,6 +70,7 @@
7070
{"slug":"qwen-image-2.0-pro","name":"Qwen-Image-2.0-Pro","description":"Qwen-Image-2.0系列满血版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。满血版具备2.0系列最强的文字渲染能力和真实质感。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-2.0-pro","name":"Qwen-Image-2.0-Pro","capabilities":["IG"]}],"detailPath":"groups/qwen-image-2.0-pro.json"}
7171
{"slug":"qwen-image-2.0","name":"Qwen-Image-2.0","description":"Qwen-Image-2.0系列加速版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。加速版有效实现了模型效果和性能的最佳平衡。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-2.0","name":"Qwen-Image-2.0","capabilities":["IG"]}],"detailPath":"groups/qwen-image-2.0.json"}
7272
{"slug":"qwen-image-3.0-pro","name":"Qwen-Image-3.0-Pro","description":"内容丰实：支持最大 4.5k token 输入，支持图中图密集信息排版，让报纸、分镜、菜单、试卷等复杂版面一次生成。\n细节真实：支持 10px 小字精准渲染，微表情、毛孔、发丝等细节生动还原，逼近真实摄影的质感。\n知识厚实：支持 12 国语言、20+ 字体原生渲染，主流网页、游戏、直播等界面仿真，外部知识全纳入。\nQwen-Image-3.0-Pro 不只是在追求\"好看\"，更在追求**“好用”**——让图像生成真正成为可落地的生产力工具。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-3.0-pro","name":"Qwen-Image-3.0-Pro","capabilities":["IG"]}],"detailPath":"groups/qwen-image-3.0-pro.json"}
73+
{"slug":"qwen-image-3.0","name":"Qwen-Image-3.0","description":"指令理解清晰：支持最大 4.5k token 输入，复杂图文指令一次生成到位。\n文字渲染稳定：10px 小字、12 国语言、20+ 字体清晰可辨，信息图和界面直接可用。\n批量产出趁手：海报、网页、界面等日常任务可批量生成，成本更优，适合持续创作。\nQwen-Image-3.0 Standard 不只在追求生成效果，更在追求日常创作里的顺手与省心——让图像生成成为可持续的内容生产力。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-3.0","name":"Qwen-Image-3.0","capabilities":["IG"]}],"detailPath":"groups/qwen-image-3.0.json"}
7374
{"slug":"qwen-image-edit-max","name":"Qwen-Image-Edit-Max","description":"千问图像编辑模型Max系列，提供更稳定、更丰富的编辑能力：提升工业设计与几何推理能力；提升角色一致性；减轻偏移问题；集成Lora能力，可以进行更多功能的图像编辑。此版本为2026年1月16日快照。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-edit-max","name":"Qwen-Image-Edit-Max","capabilities":["IG"]}],"detailPath":"groups/qwen-image-edit-max.json"}
7475
{"slug":"qwen-image-edit","name":"Qwen-Image-Edit-Plus","description":"千问系列图像编辑Plus模型，在首版Edit模型基础上进一步优化了推理性能与系统稳定性，大幅缩短图像生成与编辑的响应时间；支持单次请求返回多张图片，显著提升用户体验。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":2,"items":[{"model":"qwen-image-edit","name":"Qwen-Image-Edit","capabilities":["IG"]},{"model":"qwen-image-edit-plus","name":"Qwen-Image-Edit-Plus","capabilities":["IG"]}],"detailPath":"groups/qwen-image-edit.json"}
7576
{"slug":"qwen-image-max","name":"Qwen-Image-Max","description":"千问图像生成模型Max系列，在各类生成任务中表现出色，相较Plus系列大幅度降低生成图片的AI感，提升图像真实性；具备更真实的人物质感、更细腻的自然纹理、更美观的文字渲染。","primaryCapability":"IG","capabilities":["IG"],"providers":["qwen"],"itemCount":1,"items":[{"model":"qwen-image-max","name":"Qwen-Image-Max","capabilities":["IG"]}],"detailPath":"groups/qwen-image-max.json"}

‎skills/bailian-docs-llm-wiki/models/groups/qwen-image-3.0-pro.json‎

Lines changed: 27 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,32 @@
1818
],
1919
"provider": "qwen",
2020
"model": "qwen-image-3.0-pro",
21+
"prices": [
22+
{
23+
"priceUnit": "每张",
24+
"price": "0.02",
25+
"type": "qima_input_1k",
26+
"priceName": "1K图片输入"
27+
},
28+
{
29+
"priceUnit": "每张",
30+
"price": "0.02",
31+
"type": "qima_input_2k",
32+
"priceName": "2K图片输入"
33+
},
34+
{
35+
"priceUnit": "每张",
36+
"price": "0.25",
37+
"type": "qima_output_1k",
38+
"priceName": "1K图片生成"
39+
},
40+
{
41+
"priceUnit": "每张",
42+
"price": "0.5",
43+
"type": "qima_output_2k",
44+
"priceName": "2K图片生成"
45+
}
46+
],
2147
"qpmInfo": {
2248
"model-default-actual": {
2349
"count_limit_period": 60,
@@ -30,6 +56,7 @@
3056
"type": "model-default"
3157
}
3258
},
59+
"priceTimeBands": [],
3360
"capabilities": [
3461
"IG"
3562
],

‎skills/bailian-docs-llm-wiki/models/groups/qwen-image-3.0.json‎

Lines changed: 78 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,78 @@
1+
{
2+
"name": "Qwen-Image-3.0",
3+
"description": "指令理解清晰：支持最大 4.5k token 输入，复杂图文指令一次生成到位。\n文字渲染稳定：10px 小字、12 国语言、20+ 字体清晰可辨，信息图和界面直接可用。\n批量产出趁手：海报、网页、界面等日常任务可批量生成，成本更优，适合持续创作。\nQwen-Image-3.0 Standard 不只在追求生成效果，更在追求日常创作里的顺手与省心——让图像生成成为可持续的内容生产力。",
4+
"items": [
5+
{
6+
"inferenceMetadata": {
7+
"response_modality": [
8+
"Image"
9+
],
10+
"request_modality": [
11+
"Text",
12+
"Image"
13+
]
14+
},
15+
"description": "指令理解清晰：支持最大 4.5k token 输入，复杂图文指令一次生成到位。 文字渲染稳定：10px 小字、12 国语言、20+ 字体清晰可辨，信息图和界面直接可用。 批量产出趁手：海报、网页、界面等日常任务可批量生成，成本更优，适合持续创作。 Qwen-Image-3.0 Standard 不只在追求生成效果，更在追求日常创作里的顺手与省心——让图像生成成为可持续的内容生产力。",
16+
"features": [
17+
"model-experience"
18+
],
19+
"provider": "qwen",
20+
"model": "qwen-image-3.0",
21+
"prices": [
22+
{
23+
"priceUnit": "每张",
24+
"price": "0.02",
25+
"type": "qima_input_1k",
26+
"priceName": "1K图片输入"
27+
},
28+
{
29+
"priceUnit": "每张",
30+
"price": "0.02",
31+
"type": "qima_input_2k",
32+
"priceName": "2K图片输入"
33+
},
34+
{
35+
"priceUnit": "每张",
36+
"price": "0.18",
37+
"type": "qima_output_1k",
38+
"priceName": "1K图片生成"
39+
},
40+
{
41+
"priceUnit": "每张",
42+
"price": "0.18",
43+
"type": "qima_output_2k",
44+
"priceName": "2K图片生成"
45+
}
46+
],
47+
"qpmInfo": {
48+
"model-default-actual": {
49+
"count_limit_period": 60,
50+
"count_limit": 1,
51+
"type": "model-default"
52+
},
53+
"model-default": {
54+
"count_limit_period": 60,
55+
"count_limit": 1,
56+
"type": "model-default"
57+
}
58+
},
59+
"priceTimeBands": [],
60+
"capabilities": [
61+
"IG"
62+
],
63+
"versionTag": "MAJOR",
64+
"latestOnlineAt": "2026-08-04T15:55:59.000+00:00",
65+
"inferenceProvider": "aliyun-bailian",
66+
"name": "Qwen-Image-3.0",
67+
"category": "Visual",
68+
"samples": {
69+
"dashscope": {
70+
"default": {
71+
"curl": "curl --location 'https://[workspace-id].cn-beijing.maas.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation' \\\n--header 'Content-Type: application/json' \\\n--header \"Authorization: Bearer $DASHSCOPE_API_KEY\" \\\n--data '{\n \"model\": \"qwen-image-3.0\",\n \"input\": {\n \"messages\": [\n {\n \"role\": \"user\",\n \"content\": [\n {\n \"text\": \"画面是一张竖幅户外人像摄影，整体从上到下呈现温暖的午后街景氛围。顶部左侧到上方大面积被深绿色藤蔓和橙色小花覆盖，花叶从建筑檐口自然垂落，受阳光照射的叶片呈黄绿色高光，阴影处则偏深绿，形成浓密而柔和的背景层次。左上至中上区域是一块深蓝色横向招牌，招牌表面较暗、略带磨砂质感，上面以白色哥特体大字写着 Il Messaggero，文字位于画面左侧偏上，部分被前景花叶轻微遮挡，字体高对比、带装饰性尖角和粗细变化。招牌下方是报刊亭或书报摊的玻璃展示窗，黑色金属框架将橱窗分隔成多个矩形区域，内部陈列着许多报纸、杂志和书刊封面，但大多因景深虚化和光线反射而难以辨读，形成浅色纸张与深色边框交错的背景纹理。画面右上方是强烈的逆光区域，阳光从街道尽头照入，背景建筑被虚化成米灰色块面，边缘柔和，呈现明显的浅景深效果。画面中部偏右是一名年轻成年女性的半身至膝上人像，她回头面向镜头微笑，身体略向右转，肩背朝向观者，姿态自然放松。她有长而浓密的黑色波浪卷发，发丝被逆光勾勒出金色轮廓光，发梢在右侧向外散开，显得轻盈蓬松。她肤色白皙，脸型柔和偏鹅蛋形，眉形细致，眼睛明亮，眼妆清透，睫毛明显，面部带有自然高光，唇部为柔和珊瑚红色，笑容露齿，表情亲切明朗。她佩戴小巧耳饰，身穿黑色细肩带露背连衣裙，面料颜色深黑、轮廓简洁，细肩带从肩部向背部延伸，背部线条清晰。画面下部偏左到中部，她双手抱着一束玫瑰花，花束体积较大，主要由橙色、杏色、粉色和浅桃色玫瑰组成，花瓣层层卷曲，边缘被阳光照亮，绿色叶片和长花茎从花束下方垂出，花束与黑色裙装形成鲜明色彩对比。右侧背景是一条被阳光照亮的城市街道，地面呈暖灰与金黄色调，远处建筑、街边设施和一个模糊的红色圆形交通标志位于右下远景，均因焦外虚化而只保留色块和轮廓。整张照片采用暖色胶片感处理，带有细腻颗粒、柔和对比和明显逆光边缘光，人物位于视觉焦点，背景报刊亭、花藤、街道和阳光共同营造出浪漫、明亮、都市漫步式的氛围。\"\n }\n ]\n }\n ]\n },\n \"parameters\": {\n \"prompt_extend\": true\n }\n}'",
72+
"docUrl": "https://help.aliyun.com/document_detail/3047054.html"
73+
}
74+
}
75+
}
76+
}
77+
]
78+
}

‎skills/bailian-docs-llm-wiki/models/groups/qwen-vl-ocr.json‎

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -271,8 +271,7 @@
271271
"maxInputTokens": 30000,
272272
"offlineInfo": {
273273
"inference": {
274-
"announceUrl": "https://www.aliyun.com/notice/118434",
275-
"offlineTime": "2026-10-10 00:00:00"
274+
"announceUrl": ""
276275
}
277276
},
278277
"inferenceProvider": "aliyun-bailian",

‎skills/bailian-docs-llm-wiki/models/index.json‎

Lines changed: 22 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -1,16 +1,16 @@
11
{
2-
"updatedAt": "2026-08-04",
3-
"totalFamilies": 176,
4-
"totalModels": 394,
2+
"updatedAt": "2026-08-05",
3+
"totalFamilies": 177,
4+
"totalModels": 395,
55
"profileCoverage": {
6-
"full": 316,
7-
"partial": 34,
6+
"full": 331,
7+
"partial": 20,
88
"structured-only": 44,
99
"none": 0
1010
},
1111
"capabilityDistribution": {
1212
"TG": 36,
13-
"IG": 31,
13+
"IG": 32,
1414
"VG": 26,
1515
"TTS": 16,
1616
"Reasoning": 15,
@@ -27,7 +27,7 @@
2727
"3D-generation": 1
2828
},
2929
"providerDistribution": {
30-
"qwen": 107,
30+
"qwen": 108,
3131
"qwen-domain-model": 34,
3232
"wan": 13,
3333
"happyhorse": 4,
@@ -1226,6 +1226,21 @@
12261226
"qwen-image-3.0-pro"
12271227
]
12281228
},
1229+
{
1230+
"slug": "qwen-image-3.0",
1231+
"name": "Qwen-Image-3.0",
1232+
"primaryCapability": "IG",
1233+
"capabilities": [
1234+
"IG"
1235+
],
1236+
"providers": [
1237+
"qwen"
1238+
],
1239+
"itemCount": 1,
1240+
"items": [
1241+
"qwen-image-3.0"
1242+
]
1243+
},
12291244
{
12301245
"slug": "qwen-image-edit-max",
12311246
"name": "Qwen-Image-Edit-Max",

‎skills/bailian-docs-llm-wiki/models/index.md‎

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
# 百炼模型市场索引
22

3-
> 自动生成 · 共 176 个模型家族 · 394 个主干模型 · 更新于 2026-08-04
3+
> 自动生成 · 共 177 个模型家族 · 395 个主干模型 · 更新于 2026-08-05
44
55
**机器查询走结构化文件**：
66

@@ -92,7 +92,7 @@ join：`models.jsonl[].family == families.jsonl[].slug == index.json.families[].
9292
- [通义法睿-Plus-32K](groups/farui-plus.json) — 通义法睿是以通义千问为基座经法律行业数据和知识专门训练的法律行业大模型产品，综合运用了模型精调、强化学习、 RAG检索增强、法律Agent技术，具有回答法律问题、推理法律适用、推荐裁判类案、辅助案情分…
9393
- 模型：`farui-plus`
9494

95-
## 图像生成 `IG` — 31 个家族
95+
## 图像生成 `IG` — 32 个家族
9696

9797
- [AI试衣-Plus版](groups/aitryon-plus.json) — aitryon-plus是一款效果出众的虚拟试衣图片生成模型，可基于服饰平拍图片以及人物正面全身照，输出服饰的人物试衣效果图片。 相较于aitryon模型，aitryon-plus模型在图片清晰度、服…
9898
- 模型：`aitryon-plus`
@@ -110,6 +110,8 @@ join：`models.jsonl[].family == families.jsonl[].slug == index.json.families[].
110110
- 模型：`qwen-image-2.0`
111111
- [Qwen-Image-2.0-Pro](groups/qwen-image-2.0-pro.json) — Qwen-Image-2.0系列满血版模型，实现了图片生成和图片编辑的融合；具备更专业的文字渲染1k token指令支持能力、更细腻的真实质感，细腻刻画写实场景、更强的语义遵循能力。满血版具备2.0系…
112112
- 模型：`qwen-image-2.0-pro`
113+
- [Qwen-Image-3.0](groups/qwen-image-3.0.json) — 指令理解清晰：支持最大 4.5k token 输入，复杂图文指令一次生成到位。 文字渲染稳定：10px 小字、12 国语言、20+ 字体清晰可辨，信息图和界面直接可用。 批量产出趁手：海报、网页、界面…
114+
- 模型：`qwen-image-3.0`
113115
- [Qwen-Image-3.0-Pro](groups/qwen-image-3.0-pro.json) — 内容丰实：支持最大 4.5k token 输入，支持图中图密集信息排版，让报纸、分镜、菜单、试卷等复杂版面一次生成。 细节真实：支持 10px 小字精准渲染，微表情、毛孔、发丝等细节生动还原，逼近真实…
114116
- 模型：`qwen-image-3.0-pro`
115117
- [Qwen-Image-Edit-Max](groups/qwen-image-edit-max.json) — 千问图像编辑模型Max系列，提供更稳定、更丰富的编辑能力：提升工业设计与几何推理能力；提升角色一致性；减轻偏移问题；集成Lora能力，可以进行更多功能的图像编辑。此版本为2026年1月16日快照。

0 commit comments

Comments
 (0)