Kyeo 的 Veo3.1 Lite 等同 Google 官方 Lite 嗎？

Kyeo 的 Veo3.1 Lite 等同 Google 官方 Lite 嗎？ — 本頁描述目前已驗證的站內合約，不公開也不依賴內部供應商模型識別碼。

`REFERENCE_2_VIDEO` 目前支援幾張圖？

`REFERENCE_2_VIDEO` 目前支援幾張圖？ — 參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。

1080p 和 4K 為什麼不是基礎參數？

1080p 和 4K 為什麼不是基礎參數？ — 升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。

目前可以直接使用純文字生成嗎？

目前可以直接使用純文字生成嗎？ — 純文字、首尾影格與 1～3 張參考圖片路徑會依已提交圖片選擇對應生成模式。

參考圖片能保證角色或風格一致嗎？

參考圖片能保證角色或風格一致嗎？ — 音訊能力仍屬實驗性且不保證可用；交付前必須逐項人工檢查結果。

AI 影片模型

Google

Veo

Veo3.1 Lite

Q: 基礎生成與 Extend 各消耗多少點數？

基礎生成與 Extend 各消耗多少點數？ — 直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。

Kyeo AI 的 Veo3.1 Lite 支援純文字、首尾影格與參考圖片生成，可直接選擇 720p、1080p、4K，以及 4、6、8 秒時長。依解析度為 15 / 23 / 75 點數.

使用此模型查看點數費用回到模型廣場

能力標籤

TEXT / FRAME / REFERENCE

點數消耗

720p 15 · 1080p 23 · 4K 75

提示詞上限

5,000 字元

上傳限制

JPEG / PNG · 1–3 · 10 MB

30 秒快速瞭解

純文字、首尾影格與 1～3 張參考圖片路徑會依已提交圖片選擇對應生成模式。 TEXT_2_VIDEO / FIRST_AND_LAST_FRAMES_2_VIDEO / REFERENCE_2_VIDEO.
直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。 720p / 1080p / 4K.
參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。 REFERENCE_2_VIDEO.
音訊能力仍屬實驗性且不保證可用；交付前必須逐項人工檢查結果。 4 / 6 / 8 s.

熱門搜尋: Veo3.1 Lite 怎麼用 · Veo3.1 Lite REFERENCE_2_VIDEO 是什麼 · Veo3.1 Lite 價格

示例

Cinematic scene

A realistic 15-second vertical beauty advertisement video of a young East Asian woman with long dark brown hair in a bright modern bathroom with wooden walls, large window with blinds, and natural sunlight streaming in. She wears a simple beige ribbed tank top and has natural glowing makeup with soft pink lips. Start with her holding a frosted glass spray bottle with a gold pump and white label, smiling brightly at the camera while looking slightly to the side. On-screen text appears: “My Hair Refresh ✨”. She speaks enthusiastically: “If your hair looks dry and messy in the morning, try this: just a little serum makes such a difference!” She pumps a small amount of white serum onto her palm, then runs her fingers through her hair, applying the product. Text changes to “Smooth & Frizz-Free ✨”. She flips and smooths her hair, showing the transformation. Next, she runs both hands through her now shiny, smooth, silky hair while smiling and speaking: “Look how smooth and shiny my hair looks now!” Text changes to “Fresh Hair Finish ✨”. She continues flipping and tossing her hair playfully, looking happy and confident. Final text appears: “Try It ✨”. She holds the bottle up again, smiling at the camera, and says: “This is definitely staying in my morning routine.” Soft natural lighting, warm golden hour sunlight, high-quality realistic skin texture, cinematic beauty commercial style, smooth camera movements, vertical 9:16 format, 15 seconds duration.

Cinematic scene

CAMERA / LOOK: Handheld mini DV camcorder footage filmed by the subject herself. Slight hand shake, occasional focus hunting, imperfect framing, natural zoom adjustments, soft tape-like image quality, subtle grain, realistic auto-exposure shifts from bedroom lighting. Natural skin tones, mild motion blur, authentic consumer camcorder aesthetic rather than polished cinematic footage. STYLE: Cozy packing vlog with gentle ASMR elements. Relaxed pacing, minimal dialogue, candid moments, natural pauses. Focus on satisfying everyday sounds: suitcase zippers, clothing folds, fabric rustling, cosmetic pouch clicks, hangers sliding, bag buckles, paper travel documents, and soft footsteps around the room. SUBJECT: Young Korean woman in her 20s with long straight black hair, oversized cream hoodie, loose lounge shorts, cozy socks, minimal jewelry, natural makeup. Calm, excited energy before a short weekend trip. SETTING: Small bedroom on a Friday evening. Warm bedside lamp, open suitcase on the bed, neatly organized clothes, travel essentials on a desk, soft ambient room lighting, quiet atmosphere. STORYBOARD: (2s, propped medium shot) She places the camera on a shelf facing the bed and sits beside an open suitcase. "Packing for a little weekend escape." (2s, overhead shot) Neatly folds sweaters and places them into the suitcase. Soft fabric sounds only. (2s, close-up) Rolling clothes carefully to save space. Fabric rustling and zipper sounds. (2s, handheld closet shot) Choosing between two outfits. Holds them up and smiles. "I always bring too many clothes." (2s, detail shot) Places skincare products into a travel pouch. Small bottle clicks and pouch zipper sounds. (2s, close insert) Arranges sunglasses, charger, headphones, and a small notebook inside a side compartment. No dialogue. (2s, medium shot) Sits on the floor checking items from a handwritten packing list. "Just making sure I don't forget anything." (2s, macro shot) Passport-sized travel wallet, train ticket, and room key pouch being placed into a handbag. Paper and leather sounds only. (2s, propped shot) Closes the suitcase and gently presses down while zipping it shut. Long satisfying zipper sound. (2s, warm ending shot) Places the packed suitcase beside the door and looks back toward the camera with a relaxed smile. "I think I'm ready." (1.5s, final shot) Walks back to the camera and reaches toward the lens. "See you on the trip." Hand softly covers the lens as recording ends. AUDIO NOTES: Prioritize natural room ambience and realistic object sounds. Clothing folds, zippers, bag clasps, paper rustling, footsteps, and suitcase wheels should be clearly audible. Dialogue should be quiet and casual, as if speaking to a small personal vlog audience. REALISM NOTES: Authentic body language, natural blinking, genuine smiles, occasional pauses while thinking, realistic packing mistakes, imperfect framing, focus breathing, and everyday bedroom lighting. The footage should resemble a genuine personal weekend-trip vlog captured on a consumer camcorder, not a commercial or AI-generated production.

Cinematic scene

an athletic build of the character leaping from the skyscrapper building, while he falling down in a rapid speed, the camera focus on his face, his eyes started to glow green with a furious face he scream and then a green aura wrapped around his body as his body gradually transform larger and then green lightning struck him and then an extreme transformation of him into Kaiju no.8 begin. As he crashing down to the ground, smoke, dust and debris exploded. as the smoke and dust gradually became visible,a Masterpiece, ultra-detailed dark sci-fi manga illustration of Kaiju No. 8 in an explosive combat pose, massive armored biomechanical body with razor-sharp white horned skull helmet, glowing neon cyan energy veins pulsing through cracked obsidian exoskeleton, monstrous fanged grin, blazing luminous eyes, enormous fist thrust toward the viewer in extreme forced perspective, debris, shattered concrete and sparks erupting from the impact, swirling plasma rings, lightning arcs, heavy rain, cinematic motion blur, post-apocalyptic Tokyo battlefield, destroyed skyscrapers, smoke, fire and atmospheric fog, oversized distressed typography, Japanese kanji, military HUD interface, classified dossier graphics, barcode elements, layered mixed-media editorial composition, grunge textures, cracked ink, manga cover aesthetic, ultra-sharp linework, dramatic high-contrast lighting, volumetric effects, premium commercial poster design, mythic monster energy, hyper-detailed, 8K, vertical 9:16, masterpiece.

Cinematic scene

Main Subject: Young Chinese woman, early 20s. Natural appearance. Black hair loosely tied back with soft bangs and loose strands. Minimal makeup. Comfortable everyday summer clothing: faded cream sleeveless top, loose light-blue jeans, worn canvas sneakers. Consistent identity, clothing, and appearance throughout. Location: Authentic rural Chinese village during a warm summer afternoon. Whitewashed walls, grey tiled roofs, narrow stone paths, vegetable gardens, bamboo baskets, clay pots, citrus trees, hanging laundry, old wooden gates, chickens occasionally wandering nearby. No modern commercial elements. Camera Style (Critical) Early-2000s consumer DV camcorder. Friend or family member casually filming. Heavy handheld shake. Late reframing. Frequent autofocus hunting. Lens breathing. Exposure pumping when moving between sunlight and shade. Slight motion blur. Rolling shutter. Digital compression artifacts. Soft detail. Slightly faded colors. Low contrast. Mild sensor noise. No stabilization. No cinematic grading. 00:00–00:02 The camera follows her entering through an old wooden courtyard gate. She spots her grandmother watering vegetables and immediately smiles. The camera clips the edge of the gate slightly while entering. Autofocus struggles to switch from the gate to her face. 00:02–00:04 Grandmother waters a small vegetable garden. The young woman walks over and takes the watering can for a moment. They exchange a few casual words and laugh naturally. The camera briefly focuses on tomato plants instead of them. 00:04–00:06 Near the garden. Grandmother points toward something growing among the vegetables. The young woman crouches to inspect it. Wind moves leaves and loose hair strands. Exposure shifts as clouds pass overhead. 00:06–00:08 In the courtyard. Grandmother hands her a freshly peeled orange segment. She takes a bite and smiles. The operator catches the moment slightly late. Camera drifts off-center. 00:08–00:10 They fold laundry together beside a clothesline. Fabric moves gently in the summer breeze. The camera briefly loses focus and hunts between hanging clothes and faces. 00:10–00:12 Sitting together on small stools in the shade. Grandmother says something amusing. The young woman laughs and lightly leans toward her. The shot feels completely unplanned and natural. 00:12–00:15 The young woman leaves through the courtyard gate carrying a small bag of fruit. Halfway down the path she turns back. Her grandmother is still standing there watching. They exchange one final wave. The camera zooms slightly, loses focus for a moment, then recording cuts off abruptly mid-motion. Audio Birds. Distant village sounds. Wind through trees. Rustling leaves. Laundry fluttering. Chickens. Footsteps on stone. Quiet family conversation. No music. No narration. Goal A forgotten family camcorder recording from the early 2000s. Warm, authentic, deeply nostalgic, emotionally universal, and completely believable.

Cinematic scene

Main subject: young Japanese woman, early 20s, natural everyday appearance, faded sage-green sleeveless crop top, loose relaxed-fit cropped trousers, white canvas sneakers, thin gold chain necklace, black straight hair in a low messy bun with face-framing strands. Realistic skin texture, minimal makeup, soft and gentle personality. Maintain consistent identity, clothing, hairstyle, and appearance throughout the entire video. Location: Authentic Japanese residential neighborhood during a calm late morning. Narrow paved streets, low-rise wooden and stucco homes, small entryways with potted plants, bicycles leaning against walls, vending machine absent, utility poles, tangled overhead wires, mature trees casting moving shadows, quiet suburban atmosphere. No stores, advertisements, cafés, crowds, or commercial activity. Visual Style: Ultra-realistic documentary realism. Genuine candid behavior. Natural body language. Unscripted slice-of-life feeling. Strong environmental authenticity. Rich real-world details and believable human motion. Camera Style: Early-2000s consumer DV camcorder aesthetic. Friend casually recording everyday moments. Heavy handheld shake, imperfect framing, frequent autofocus hunting, lens breathing, exposure pumping when moving between sun and shade, occasional motion blur, subtle rolling shutter, mild digital compression artifacts, faded colors, soft contrast, slight sensor noise. No stabilization. No cinematic camera moves. No modern color grading. **00:00–00:02** Outside a small house entrance. She sits on a low wooden step tying her sneaker laces. A light breeze moves loose strands of hair. She smiles naturally while the camera struggles to hold focus. **00:02–00:04** The camera follows her into a narrow street lined with potted plants and low walls. She notices a stray cat approaching and crouches down. Framing drifts off-center as the operator tries to keep up. **00:04–00:06** She gently pets and feeds the cat. Autofocus repeatedly shifts between her face and the animal. Morning sunlight flickers through leaves overhead. **00:06–00:08** Small front yard beside her house. She hangs laundry on a balcony rack while fabrics sway in the breeze. Exposure changes as clouds briefly pass overhead. **00:08–00:10** On a quiet engawa (wooden porch) with a ceramic tea cup. She sits comfortably watching the neighborhood, occasionally brushing hair behind her ear. Loose handheld side angle with natural camera drift. **00:10–00:12** Close side profile. Someone off-camera greets her. She turns, raises her hand, smiles warmly, and casually says, "Ohayou." The camera catches the moment slightly late. **00:12–00:15** Walking slowly down a tree-lined residential lane holding her tea cup. She notices the camera, gives a small genuine smile, then looks away and continues walking. Recording cuts abruptly to black mid-motion as if the camcorder was switched off. Audio: Natural ambient sound only — morning birds, distant scooters, light wind, leaves rustling, faint neighborhood chatter, cat sounds, footsteps on pavement, fabric moving on a balcony rack, subtle residential ambience. No music. No sound design. No narration. Goal: Authentic Japanese neighborhood life captured like a forgotten home video from the early 2000s — candid, imperfect, realistic, warm, and deeply believable.

Cinematic scene

THE LAST TRAIN Duration: 15 SecondsAspect Ratio: 16:9 Genre: Supernatural Psychological ThrillerVisual Style: Photorealistic cinematic realism, cold blue grading, atmospheric lighting, subtle tension, natural film grain, realistic physics. GLOBAL CONTINUITY Use the attached reference image @image1 as the permanent facial anchor. Preserve exact face shape, eyes, skin tone, age, and expression in every shot. No exaggeration—emotion comes from lighting, posture, and movement only. Wardrobe evolves naturally. She starts in a charcoal wool coat and dark boots with neatly styled hair. During the escape, the coat opens, revealing a silver crescent necklace. Hair transitions from neat to windblown. Hair color never changes. All shots use a consistent 35–50 mm cinematic perspective with subtle lens breathing, natural focus, realistic cloth and hair physics, drifting fog, and cold blue lighting. CUT 1 | 0:00–0:03 | ISOLATION Wide shot of an empty railway platform in cold fog under flickering lights. She stands alone in a charcoal coat, checking her phone as signal drops to zero. She exhales visible breath. Slow dolly-in with soft focus pull. Fog drifts naturally. Cut as she lowers her phone. CUT 2 | 0:03–0:06 | DISCOVERY Close-up of a cracked station clock frozen at 11:47. Her faint reflection appears in the glass. Slow push-in. Cold blue lighting remains consistent. The second hand suddenly jerks. Cut on the movement. CUT 3 | 0:06–0:09 | AWAKENING Clock jerks violently, cracks deepen. Camera pulls back to reveal the Conductor standing behind her. He wears a dark vintage uniform, face hidden in shadow, faint pocket watch glow. Wind lifts strands of her hair as she begins turning. Lights flicker colder. Cut on strongest flicker. CUT 4 | 0:09–0:12 | ESCAPE Handheld tracking shot as she runs toward the tunnel. Her coat opens, revealing a silver crescent necklace. Hair flows in wind, breath visible. She glances back once. Tunnel lights flicker on sequentially. Cut on whip-pan. CUT 5 | 0:12–0:15 | LOOP Ground-level wide shot. Shadow tendrils erupt from the clock, pulling her and the Conductor inside. They dissolve into interference. Necklace remains visible until the final moment. Camera pulls back to a static shot. The clock reads 11:47 again. A train rushes past. Lights fade until only the clock remains. Final tick → hard cut to black. SOUND DESIGN Continuous cinematic score.Start with ambient drone and wind → add distant railway sounds → soft piano.Heartbeat pulse builds with clock ticks and strings.During escape, rhythm accelerates with percussion and tension. At capture, a strong bass impact.Afterward, only faint interference and wind. Final moment: silence, then one clean clock tick at 11:47. GLOBAL NEGATIVE PROMPT No identity drift, no face distortion, no extra limbs, no duplicate characters, no flickering, no texture issues, no unrealistic motion, no clipping, no artifacts.

一句話瞭解

這個模型表現如何？

純文字、首尾影格與 1～3 張參考圖片路徑會依已提交圖片選擇對應生成模式。直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。

它擅長什麼: 目前定位是 Kyeo 的 Lite 公開資料接入頁，而不是 Google 官方 Veo 3.1 Lite 的完整鏡像。
適合哪些使用者: 適合用非關鍵素材逐一驗證文字、首尾影格或參考圖片路徑，並在基礎任務成功後另行評估結果操作。
為什麼要在 Kyeo AI 使用: Kyeo AI 將 Veo 的可用參數、點數消耗與結果操作放在一起，方便比較方案並立即開始創作。

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 影片模型

廠商

Google

模型家族

Veo

整合方式

目前站內整合

執行方式

非同步任務

提示詞上限

5,000 字元

上傳限制

JPEG / PNG · 1–3 · 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。

Veo3.1 Lite

Google Veo

Veo Video

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。

Veo3.1 Lite 怎麼用

Veo3.1 Lite REFERENCE_2_VIDEO 是什麼

Veo3.1 Lite 價格

Veo3.1 Lite 和 Quality 差異

Veo3.1 Lite 和 Lite 差異

Veo3.1 Lite 支援幾張參考圖

Veo3.1 Lite Extend 是什麼

Veo3.1 Lite Get 1080P 是什麼

Veo3.1 Lite Get 4K 是什麼

Kyeo AI 的 Veo3.1 Lite 與 Google 穩定版 Lite 有什麼區別？

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。

什麼情況可以先測試

當你需要驗證 Kyeo 目前 Lite 流程中的一條明確路徑，並願意先用非關鍵素材確認請求、回呼、結果與扣點時，可以從最小案例開始。

什麼時候先比較其他模型

若只需要較低的 Kyeo AI 扣點，可比較 Lite；Sora 2 Pro 已屬 Legacy，不適合作為預設交付基準。

最小驗證路徑

最小驗證路徑: 交付前請以非敏感樣例核對生成模式、時長、實際像素、音訊、網址期限與實際扣點。

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。

Veo3.1 Lite 和 Veo3.1 Lite 差異

Veo3.1 Lite vs Veo3.1 Lite

Veo3.1 Lite 和 Veo3.1 Quality 差異

Veo3.1 Lite vs Veo3.1 Quality

Veo3.1 Lite 和 Sora 2 Pro 差異

Veo3.1 Lite vs Sora 2 Pro

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。

Kyeo 目前角色

Veo3.1 Lite

包含三種輸入路徑與結果操作的站內 Veo 工作流程

Veo3.1 Quality

高品質檔

生成模式

Veo3.1 Lite

文字 + 首尾幀 + 參考圖

Veo3.1 Quality

基礎文生 / 圖生，不開參考圖模式

結果操作入口

Veo3.1 Lite

成功任務後顯示 Extend / 1080p / 4K

Veo3.1 Quality

同樣支援結果動作

本地基礎扣點

Veo3.1 Lite

固定 15 點數

Veo3.1 Quality

固定 225 點數

目前驗證重點

Veo3.1 Lite

先確認一條基礎路徑與任務狀態閉環

Veo3.1 Quality

目前驗證重點 — Veo3.1 Quality: 目前可用的同類模型，可從詳情頁繼續比較功能與價格。

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。

Veo：快速試用亮點

Veo：快速試用亮點: 本頁描述目前已驗證的站內合約，不公開也不依賴內部供應商模型識別碼。

基礎請求與結果操作必須分開計價

基礎請求與結果操作必須分開計價: 直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。

功能詳解

目前站內的三種輸入路徑

目前站內的三種輸入路徑: 純文字、首尾影格與 1～3 張參考圖片路徑會依已提交圖片選擇對應生成模式。

參考圖片模式是 Kyeo 的產品選擇

參考圖片模式是 Kyeo 的產品選擇: 參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。

結果操作依賴站內原始任務

結果操作依賴站內原始任務: 升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。

適合做什麼

Veo：日常創作

Veo：日常創作: 交付前請以非敏感樣例核對生成模式、時長、實際像素、音訊、網址期限與實際扣點。

參考圖驅動影片

參考圖驅動影片: 參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。

成功任務的結果操作驗證

成功任務的結果操作驗證: 升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。

提示詞技巧

先選對 `generationType`，再最佳化 prompt

Veo Lite 目前最重要的不是多寫幾個風格詞，而是先判斷你要文字、首尾幀，還是參考圖模式。路徑選錯了，prompt 再長也很難補回來。

參考圖模式先把圖片職責分清楚

使用 1～3 張參考圖時，先標註每張圖負責主體、造型或風格中的哪一項，再逐項檢查是否混用或遺漏。

把升頻與延長視為獨立步驟

兩者都是成功任務後的獨立操作與額外扣點。解析度提高不代表內容、動作或細節一定更好。

選擇它的理由

純文字、首尾影格與 1～3 張參考圖片路徑會依已提交圖片選擇對應生成模式。 4 / 6 / 8 s.

參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。 1–3 JPEG / PNG.

直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。 15 / 23 / 75.

升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。 720p / Extend.

使用前須知

參考圖片模式接受 1～3 張經簽章驗證的 JPEG 或 PNG，只支援 16:9／9:16，且時長必須為 8 秒。 Auto.

升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。 1080p / 4K.

音訊能力仍屬實驗性且不保證可用；交付前必須逐項人工檢查結果。升解析度操作只接受本人成功的 720p 基礎任務或 Extend 結果；直接生成的 1080p、4K 結果不能再次要求升解析度。

交付前請以非敏感樣例核對生成模式、時長、實際像素、音訊、網址期限與實際扣點。 10 MB.

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。

生成模式

可選

參數類型: select

依已提交圖片決定

TEXT_2_VIDEO

FIRST_AND_LAST_FRAMES_2_VIDEO

REFERENCE_2_VIDEO

寬高比

可選

參數類型: select

預設值: 16:9

橫向 (16:9)

直向 (9:16)

自動裁切 (Auto)

解析度

可選

參數類型: select

720p

720p — 15

1080p — 23

4K — 75

時長

可選

參數類型: select

8 s

4 s

6 s

8 s

浮水印文字

可選

參數類型: text

提示詞翻譯

可選

參數類型: select

預設值: true

自動翻譯為英文

保留原文

點數消耗

依解析度為 15 / 23 / 75 點數

直接生成 720p 扣 15 點數，1080p 扣 23 點數（22.5 無條件進位），4K 扣 75 點數。Extend 扣 15；符合條件的 1080p、4K 結果操作另扣 8、60 點數。

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。

Veo3.1 Fast

Kyeo AI 的 Veo3.1 Fast 支援純文字、首尾影格與參考圖片生成，可直接選擇 720p、1080p、4K，以及 4、6、8 秒時長。依解析度為 30 / 38 / 90 點數.

AI 影片模型

依解析度為 30 / 38 / 90 點數

查看詳情立即使用

Veo3.1 Quality

Veo3.1 Quality 是 Kyeo 目前採用 Veo 3.1 Quality 公開資料識別碼的 Veo 路線，純文字或 1～2 張首尾影格圖片的基礎請求固定 225 點數；成功後可另行發起 Extend、1080p 或 4K。Quality 是工作流程名稱，不代表每次結果都能達到固定的交付品質。

AI 影片模型

依解析度為 225 / 233 / 285 點數

查看詳情立即使用