跳至主要內容
Kyeo AI
AI 圖片模型
ByteDance
Seedream

Seedream 5.0 Pro

Seedream 5.0 Pro 將文字生成與最多 10 張參考圖的編輯整合在同一入口。basic 對應 1K，high 對應 2K；編輯任務會依輸入圖片數量增加預授權，適合需要在送出前確認成本的多圖工作流程。

使用此模型查看點數費用回到模型廣場
工作流程
文字生成圖片、多圖編輯
輸出層級
1K／2K
參考圖
最多 10 張
點數
7–19
30 秒快速瞭解
  • 沒有圖片時為文字生成圖片，上傳圖片後自動進入圖片編輯
  • 8 種長寬比，basic 為 1K，high 為 2K
  • 最多 10 張 JPEG／PNG／WebP，每張不超過 10 MB
  • 從第 2 張輸入圖片起，每張增加 0.5 credits，預授權一律無條件進位

熱門搜尋: Seedream 5.0 Pro 怎麼用 · Seedream 5.0 Pro 多少錢 · Seedream 5.0 Pro 支援幾張參考圖

示例

Visual concept

Visual concept

使用上传的人物照片作为主体参考，保留人物五官比例、脸型轮廓、眼神气质、发型特征与整体辨识度，不要生成陌生人脸。生成一组高端白棚时尚写真海报，成年女性模特，黑色微卷长发，清冷眼神，皮肤真实有细节，表情自然不僵硬。 整体风格：极简白色摄影棚、时尚杂志大片、真人小手办展示感、干净高级、轻微胶片质感。人物穿黑色宽松皮衣外套、白色透感层次上衣、黑色百褶短裙、白色中筒袜、黑色厚底长靴，搭配珍珠项链和银色饰品。 场景道具：超大收据地毯、黑白棋盘格抱枕、黑色小包、摄影棚支架、白色背景。不要使用真实品牌 Logo，可使用虚构摄影棚标识和虚构收据文字。 动作参考：分别生成不同姿势，坐姿、跪姿、趴姿、侧身回头、手拿场记板、躺在收据地毯上、近距离面部特写。每张图都要符合参考动作和表情，眼神有情绪，画面高清，人物五官清晰。 英文排版：加入赛博故障风英文大字，黑白为主，带少量 RGB 偏移、扫描线、信号错位、噪点 and 细小英文注释。排版风格像独立杂志封面，不要遮挡人物脸部。 可用英文标题： STATIC MUSE PIXEL FEVER WHITE NOISE OFFLINE DESIRE BROKEN LULLABY GLITCH ANGEL ERROR ROMANCE CUT THE SIGNAL 小字文案： Beauty is not perfect. It’s real when the signal breaks. This isn’t a portrait. It’s a frequency. A moment between noise and silence. Beauty is data. Emotion is signal. You are the glitch. 高清写实摄影，真实皮肤纹理，柔和棚拍光，高级商业摄影，editorial fashion poster，clean white studio，glitch typography，RGB shift，magazine layout，gpt-image2 style。

Visual concept

Visual concept

Create a high-end studio fashion portrait inspired by the uploaded blue-background reference photograph. Recreate the original composition with matching framing, focal length feel, camera position, subject scale, perspective, and overall visual balance while maintaining a realistic photographic appearance. Replace the reference model with the character from the supplied character image. Preserve the character's authentic facial identity, bone structure, skin tone, eye shape, hairstyle, proportions, and recognizable features. Avoid beautification that alters identity. The face should remain naturally expressive and true to the source character. Match the original body language closely: upper torso visible, shoulders relaxed, chin slightly elevated, eyes directed upward beyond the camera, conveying quiet confidence and contemplation. Maintain the same sense of presence and spatial positioning within the frame. Retain the rich cobalt-blue studio backdrop and the refined editorial atmosphere. Reproduce the original lighting setup with soft directional illumination, subtle facial highlights, realistic reflections on eyewear and jewelry, and smooth shadow transitions around the jawline, neck, and clothing folds. Preserve depth, dimensionality, and studio realism. Style the character in contemporary oversized street-fashion apparel. Use a premium relaxed-fit knit sweater or sweatshirt with natural fabric draping, layered with a minimalist chain necklace and lightly tinted translucent eyewear. The wardrobe may be customized to suit the character, but should maintain the same luxurious oversized silhouette and fashion-campaign aesthetic. Keep the image tightly composed with the same portrait crop and subject placement. Ensure identical camera-to-subject distance and visual weight in the frame. The final result should feel indistinguishable from a genuine professional photoshoot captured in this exact studio environment. Photorealistic editorial photography, luxury fashion campaign, cinematic studio portrait, premium streetwear styling, deep monochromatic blue backdrop, realistic skin texture, subtle film-grade contrast, detailed fabric rendering, natural lighting falloff, commercial advertising quality, shallow depth of field, magazine-cover aesthetics. Negative Prompt: Do not alter the background color or studio environment. Do not change the pose, framing, crop, perspective, subject distance, lighting direction, or gaze. Do not modify facial identity, facial proportions, skin tone, or hairstyle. Avoid excessive retouching, artificial skin smoothing, stylized illustration effects, exaggerated sharpening, extra accessories, text overlays, logos, watermarks, duplicate subjects, or unrealistic clothing physics.

Visual concept

Visual concept

Ultra-realistic IMAX-level Netflix-style cinematic fashion lifestyle poster portrait, 9:16 vertical composition, use the uploaded image as the primary visual reference for pose, mood, outdoor café environment and stylish attitude while redesigning the subject as a beautiful young adult woman, create a full-body shot of a confident girl sitting in a relaxed cross-legged pose on a modern wooden café bench during warm golden-hour sunlight, her luminous fair porcelain milky-white skin rendered with an ultra-smooth clean texture and perfect face-to-body skin tone consistency, no dullness and no tanning, her long voluminous dark-brown hair styled in soft lifted waves with a few face-framing strands, wearing oversized black square sunglasses, a fitted sleeveless beige ribbed tank top, black ripped shorts on the bottom, and stylish casual footwear, with a smartwatch on one wrist, a delicate bracelet and minimal elegant accessories, one arm resting casually along the bench while the other hand rests naturally near her lap, an iced coffee cup placed beside her on the table in the foreground, her facial expression showing calm confidence, soft feminine boldness and effortless attraction, with relaxed brows, slightly lifted chin, softly defined lips and a composed cool-girl attitude, camera pulled back enough to capture the entire body and full styling clearly, background featuring a premium café setting with glass windows, soft reflections, outdoor seating, modern architecture and warm sunlit ambience, sunlight falling beautifully across her face, shoulders, arms and legs with cinematic highlights and gentle shadow patterns, add stylish café-fashion overlay text integrated naturally into the poster design: large elegant headline “COFFEE & CONFIDENCE”, small side text “SUNLIT MOOD”, handwritten caption “cool girl energy”, minimal label-style text “GOLDEN HOUR EDIT”, tiny sticker text “SIP. SIT. SLAY.” and a clean bottom tagline “EFFORTLESS STYLE, EVERYDAY LUXURY”, use mixed typography styles including bold editorial lettering, soft handwritten script, minimal sticker labels and thin fashion-magazine text, placed around the subject without covering the face, sunglasses or body pose, balanced full poster layout with no empty awkward space, ultra-rich color grading with warm amber sunlight, soft skin glow, beige neutrals, deep black shorts contrast, café-brown tones and luxurious lifestyle atmosphere, captured with a premium portrait-fashion lens, shallow depth of field, HDR lighting, global illumination, photorealistic rendering, masterpiece quality, ultra-detailed textures, 8K production detail. Negative Prompt: bad anatomy, extra fingers, distorted legs, cropped body, blurry face, dull skin, uneven skin tone, flat lighting, messy shorts, unreadable text, misspelled words, poor typography, cartoon, anime, CGI look, watermark, logo, low quality.

Visual concept

Visual concept

A photorealistic fashion upper body shot of [Character Description and expression] wearing [Detailed Outfit Description]. The subject is framed inside a central white Instagram-style post border. Composition & Spacing: The white frame is perfectly centered in the middle of the image, leaving balanced empty [Soft Background Color matching the outfit] space above and below the frame to match the theme. Frame Details: The top, left, and right white borders are very thin. The bottom white section is thicker to include UI elements. The bottom section features a red heart icon, comment bubble, share icon, and bookmark icon. Text Details: Clearly visible text on the bottom panel: "[Number] likes", username "[Username]", caption "[Caption text]... more", and below that, "View all comments". 3D Pop-Out Effect & Hand Pose: [CHOOSE ONE OF THE FOLLOWING ACTIONS OR DESCRIBE YOUR OWN ACTION:] 1. (Grip): head and arms are physically popping out OVER the top and side thin borders. hands are realistically gripping the outer edges of the thin side borders, appearing to hold the frame firmly. 2. (Kiss): Her body is leaning forward towards the camera, creating a strong sense of depth. Her head remains securely framed WITHIN the top thin white border. Only her right hand, raised near her mouth in a blowing-kiss gesture, physically extends forward and breaks OUT of the thin side frame boundary. Her left hand realistically grips the side edge. 3. (Resting): She is looking directly at the camera with a gentle smile. Her head and shoulders physically pop out OVER the top thin white border. Crucially, her hands are realistically resting comfortably on top of the thicker bottom white UI panel, appearing to lean casually on the frame edge. 4. (Dynamic Frame): Her upper body and head remain secure WITHIN the boundaries of the white frame. However, both of her arms are fully extended outward dynamically, breaking past the thin borders up to her elbows, reaching straight toward the camera lens. She forms a distinct “L” shape with both hands, as if framing a scene. Due to perspective, her hands are positioned significantly closer to the camera than her face, showing strong spatial depth. Subject Dimensionality & Texture: The subject must look like a solid, three-dimensional figure, not flat. Emphasize the detailed texture of the [Specific Fabric Name from outfit description], the contours of the face, and realistic hair details. She should have distinct depth separating her from the flat [Background Color] background. Lighting & Shadows: Professional studio lighting creating soft, realistic drop shadows cast by the physical frame and, crucially, by the [popping-out elements, e.g., hands/head/arms] onto the solid flat [Background Color] background, enhancing the dramatic depth and illusion. High definition, commercial photography.

Visual concept

Visual concept

{ "subject": { "primary": "Cosmetic spray bottle", "secondary": "Hummingbird", "brand": "IDUN MINERALS", "product_name": "READY SET REFRESH" }, "packaging": { "form_factor": "Cylindrical bottle", "component_parts": [ "Main body", "Cap (flush with body)" ], "material_appearance": "Matte finish, likely plastic or coated metal", "condition": "Pristine, new", "reflections": "Vertical highlight on the left side, mirrored reflection in water below" }, "label_design": { "type": "Direct print onto bottle", "layout_orientation": "Mixed (Horizontal brand/details, Vertical product name)", "typography": { "font_family": "Sans-serif", "casing": "All uppercase", "style": "Clean, minimalist", "alignment": "Left aligned relative to the vertical axis for product name, centered horizontally for brand" } }, "text_visible": { "brand_header": [ "IDUN", "MINERALS" ], "body_vertical": [ "READY SET REFRESH", "SETTING SPRAY" ], "footer_details": [ "PRIMES THE SKIN", "SETS MAKEUP & REFRESHES", "FORMULATED WITH", "3% NIACINAMIDE" ], "reflection_text": "Inverted mirror image of all text visible on the water surface" }, "colors": { "packaging_base": "Deep red / Burgundy", "text": "White", "foliage": [ "Deep emerald green", "Forest green", "Moss green" ], "hummingbird": [ "Iridescent green (head/back)", "Dark grey/brown (wings)", "Reddish/pink (tail feathers)", "Black (beak)" ], "water": "Dark, almost black with reflections" }, "materials": { "bottle": "Matte rigid material", "environment": [ "Liquid water", "Moss", "Fern leaves", "Palm-like leaves" ], "biological": "Feathers (bird)" }, "human_elements": { "visible": false }, "environment": { "setting": "Simulated rainforest or jungle floor", "foreground": "Still water surface", "midground": "Mossy bank, bottle placement", "background": "Dense tropical foliage" }, "lighting": { "quality": "Soft, moody, atmospheric", "direction": "Side lighting from left (creating highlight on bottle)", "shadows": "Soft shadows in the foliage, dark water providing high contrast reflection", "highlights": "specular highlight on hummingbird head, diffuse highlight on bottle curve" }, "camera": { "angle": "Eye-level with water surface", "framing": "Vertical portrait", "focus": "Sharp focus on bottle and hummingbird", "depth_of_field": "Shallow (background leaves are blurred/bokeh)", "composition": "Rule of thirds (bottle on left third line, bird entering from right)" }, "style": { "category": "Product photography", "theme": "Nature, Organic, Eco-luxury", "mood": "Serene, fresh, premium" } }

Visual concept

Visual concept

[BRAND]. A high fashion campaign poster. A model with a strong expressive face as the centerpiece, wearing a signature outfit in the brand's exact official color palette. Dramatic surreal atmospheric background. Layered collage of torn-edge material swatches surrounding the model. Brand's iconic symbol woven into the scene. Logo overlaid in the center in clean serif font. Macro detail on fabric texture and stitching. Cinematic lighting, rich contrast, 8K

一句話瞭解

這個模型表現如何？

未上傳圖片時，任務會以文字生成圖片處理；上傳 1–10 張圖片後則切換為圖片編輯。兩種模式共用 8 種長寬比、basic／high、PNG／JPEG 與內容檢查控制。內容檢查只是輔助開關，不代表輸出已通過著作權、隱私、安全或事實查核。

它擅長什麼
適合需要可控輸出層級與多圖參考的圖片生成及編輯任務，重點是在同一入口切換模式並清楚計費。
適合哪些使用者
適合商品視覺、海報、角色設定、版面變體，以及需要同時參考主體、構圖和風格素材的設計流程。
為什麼要在 Kyeo AI 使用
Kyeo 會在送出前依輸出層級與參考圖數量計算整數預授權，並在任務完成後讀取實際消耗；請求欄位只依目前合約白名單送出。

關鍵參數總覽

協助你快速評估此模型是否符合你的業務情境。

模型類別
AI 圖片模型
廠商
ByteDance
模型家族
Seedream
執行方式
非同步任務
提示詞
去除前後空白後 3–5,000 個字元
輸入圖片
JPEG／PNG／WebP，每張不超過 10 MB

其他名稱

同一個模型在官方文件、社群討論與 API 串接層可能有不同稱呼，這裡幫你統一整理。

Seedream 5 Pro
Seedream 5.0 Pro 文字生成圖片
Seedream 5.0 Pro 圖片編輯

使用者常見問題

從實際任務、輸入條件和結果要求出發，整理選擇模型時最需要確認的問題。

Seedream 5.0 Pro 怎麼用
Seedream 5.0 Pro 多少錢
Seedream 5.0 Pro 支援幾張參考圖
Seedream 5.0 Pro basic 和 high 有何差異
Seedream 5.0 Pro 支援哪些圖片比例
Seedream 5.0 Pro 和 Seedream 5.0 Lite 有何差異

選擇指南

還在猶豫該用哪個模型嗎？先查看這幾個關鍵決策點。

1
多圖參考可先測試 Pro

當主體、風格、構圖或配色分別來自不同素材時，可先指定每張圖片的用途，再以同一提示詞測試組合效果。

2
需要更高標示解析度時先比較

目前 Pro 最高為 2K；如果任務最重視 3K／4K 層級，可用 Seedream 5.0 Lite 或 Seedream 4.5 進行同題比較。

3
將圖片數量納入預算

high 加上 10 張參考圖時，預授權為 19 點數；移除非必要的參考圖可以降低成本，也能減少互相衝突的視覺訊號。

使用者常搜尋的比較

把常見候選放在同一任務下比較，協助你確認輸入方式、控制項和成本差異。

Seedream 5.0 Pro 與 Seedream 5.0 Lite 的解析度和參考圖上限有何不同？
Seedream 5.0 Pro 與 Seedream 4.5 的計費方式有何不同？
Seedream 5.0 Pro 適合單圖編輯還是多圖參考？

模型比較表

還在猶豫該選哪一個嗎？這張表協助你快速看清目前模型與替代方案的核心差異。

輸出層級
Seedream 5.0 Pro
basic 1K；high 2K
Seedream 5.0 Lite
文字生成 high 3K；編輯 high 4K
Seedream 4.5
basic 2K；high 4K
參考圖上限
Seedream 5.0 Pro
最多 10 張
Seedream 5.0 Lite
最多 14 張
Qwen Image 3.0 Pro
最多 3 張
本地預授權
Seedream 5.0 Pro
7–19 點數，依層級和圖片數量而定
Seedream 5.0 Lite
固定 6 點數
Seedream 4.5
固定 7 點數

實際使用體驗

結合公開資料、目前站內邊界和典型任務，說明模型適合與不適合的情況。

Pro 名稱不代表每個面向都更高

目前可驗證的差異是工作流程、參數、輸入上限與價格。畫質、速度與指令遵循度仍應以實際任務樣本判斷。

多圖能力的價值取決於素材分工

單純堆疊多張相似圖片容易增加衝突；明確指定每張參考圖負責的元素，比只增加數量更有價值。

功能詳解

文字生成與圖片編輯分流

系統依是否有參考圖選擇相應工作流程，不會把圖片欄位送到純文字任務。

層級與格式控制

可選 basic／1K 和 high／2K 兩個層級，輸出格式為 PNG 或 JPEG。

依圖片數量預授權

基本價格由 quality 決定，編輯任務從第二張輸入圖片起加價，建立任務前就完成保守預授權。

適合做什麼

商品視覺變體

分別提供商品、環境與風格參考，說明必須保留的品牌元素，再生成不同橫式與直式素材。

角色一致性探索

用少量高品質角色參考固定服裝、髮型和配色，並在提示詞中說明可變動的姿勢與場景。

海報與版面草稿

固定畫面比例和資訊層級，生成前先確認文字內容仍需人工排版與校對。

提示詞技巧

標明每張圖片的用途

說明哪張負責主體、哪張負責風格、哪張負責構圖，避免模型自行混合所有特徵。

先 basic 再 high

先用 1K 驗證構圖與指令，再以 2K 處理選定的方向，降低高層級的試錯成本。

寫清楚保留與修改項目

編輯任務同時列出必須保留及允許改動的元素，比只描述目標效果更容易驗收。

選擇它的理由

同一入口涵蓋文字生成與最多 10 張參考圖編輯。
輸出層級、長寬比與格式限制清楚。
預授權會反映參考圖數量，不必等任務結束才發現成本改變。
輸入格式和單張圖片大小會在上傳時依模型合約驗證。

使用前須知

目前最高層級為 2K，不應宣稱為 3K 或 4K。
參考圖越多不代表結果越穩定，素材衝突可能增加重做成本。
內容檢查開關不能取代人工著作權、隱私與安全審核。
模型名稱不能當作速度、品質或商業可用性的保證。

可調整參數

協助你快速評估此模型是否符合你的業務情境。

長寬比
選填
參數類型: 下拉選單
預設：1:1
1:1
4:3
3:4
16:9
9:16
2:3
3:2
21:9
8 項
輸出品質
選填
參數類型: 下拉選單
預設：basic（1K）
basic（1K）
high（2K）
輸出格式
選填
參數類型: 下拉選單
預設：PNG
PNG
JPEG
內容檢查
選填
參數類型: 開關
預設：關閉
關閉
開啟

點數消耗

7–19 點數

basic／1K 的基本價格為 7 credits，high／2K 為 14 credits；圖片編輯從第 2 張輸入圖片起，每張另計 0.5 credits。Kyeo 會將請求總價無條件進位，因此合法請求的預授權為 7–19 點數，完成後再依實際消耗結算。

省點數小技巧

批次生成圖片或影片前，建議先用同一組素材，在目前模型與同類候選模型之間進行 A/B 測試；確認效果後再批次生成，避免浪費點數。

常見問題

同類模型推薦

還沒決定該用哪個嗎？順便比較一下這幾個同類候選模型。

Seedream 5.0 Lite

Seedream 5.0 Lite 是字節跳動較輕量的圖片模型，支援 8 種長寬比和最多 14 張參考圖片。可選 Basic 2K、High 3K 或 Ultra 4K，輸出 PNG 或 JPEG，並包含敏感內容檢查。

AI 圖片模型
6 點數 / 次
查看詳情立即使用

Seedream 4.5

Seedream 4.5 / 能力標籤 — ByteDance 官方資料著重展示 Seedream 4.5 的多圖編輯、參考細節保留、版面與文字呈現。Kyeo 的實際提供另有明確限制：提示詞最多 3,000 個字元，提供 8 種長寬比，編輯任務最多接收 14 張 JPG、PNG 或 WebP 圖片，每張不超過 10 MB。 Kyeo AI 目前每次請求固定消耗 7 點數，basic 與 high 同價。實際預算還要納入重試次數、返工率與最終可用率。

AI 圖片模型
7 點數 / 次
查看詳情立即使用

Qwen Image 3.0 Pro

Qwen Image 3.0 Pro 提供文字生成圖片與最多 3 張參考圖的圖片生成圖片，可選 1K／2K、8 種畫布尺寸、PNG／JPEG 輸出，以及提示詞擴充、負面提示詞和 seed。它會依解析度及輸入圖片數量預授權 7–14 點數。

AI 圖片模型
7–14 點數
查看詳情立即使用

資訊來源

頁面內容依據廠商資料、模型系列說明與 Kyeo AI 平台串接設定整理。

來源說明

本頁依目前站內介面合約與價格證據複核，只說明 Kyeo 已開放的文字生成、參考圖編輯、輸出層級、格式與內容檢查控制。模型廠商可能提供的其他能力，不代表本站也已提供。

最後更新: 2026-08-11
Seedream 5 Pro 目前介面資料
串接

用於核對本站目前接入版本的輸入模式、可見設定、素材限制與可用狀態；實際選項以頁面和工作台為準。