Alibaba AI API models
Alibaba AI models on MuAPI
Explore Alibaba models for chat, code, image and video generation, including Gemini, Nano Banana and Veo-style workflows available through MuAPI.
All models
90 ModelsVideo Generation Models
$0.700 / second
happy-horse-1.1-text-to-video-720p
Happy Horse 1.1 Text to Video (720p) — generate expressive 720p video clips from text prompts with vivid character motion.
$0.500 / second
wan3.0-text-to-video
Wan 3.0 Text to Video generates a video with synchronized audio directly from a text prompt, with selectable resolution, aspect ratio, and duration.
$0.700 / second
wan3.0-prime-reference-to-video
Wan 3.0 Prime Reference to Video generates a higher-fidelity video guided by a prompt plus up to 10 reference images, 5 reference videos, and 5 reference audios for visual and audio coherence.
$0.300 / second
wan2.2-image-to-video
Wan 2.2’s I2V mode brings static visuals to life with vivid, expressive animations. It interprets motion, emotion, and background dynamics from a single image to generate smooth and cinematic short videos.
$0.300 / second
wan2.1-text-to-video
WAN 2.1 turns your written prompts into vivid, cinematic video clips. Ideal for storytelling, content creation, and visualizing abstract ideas, it supports detailed natural scenes, character motion, and dramatic camera movements — all from just text.
$0.350 / second
wan2.2-animate
Wan2.2 Animate is a video-to-video model for animating a character or replacing a character in existing video clips. It replicates holistic movement and facial expressions from a reference video or pose while preserving the target character’s appearance. You upload both an image (for the character) and a video containing motion/expression, and the model generates a video where the character in your image moves like the reference. Supports 480p or 720p, up to 120 seconds
$0.300 / second
wan2.2-text-to-video
Wan 2.2’s T2V mode transforms descriptive text prompts into high-quality, stylized video sequences. It excels at generating anime-style or cinematic visuals with smooth motion and strong thematic consistency.
$0.550 / second
wan3.0-spicy-text-to-video
Wan 3.0 Spicy Text to Video generates a bold, high-contrast, high-motion video with synchronized audio directly from a text prompt, with selectable resolution, aspect ratio, and duration.
$0.100 / second
wan2.1-reference-video
WAN 2.1 is an advanced AI model that transforms one or more reference images into a coherent, animated video. By combining characters, objects, or environments from multiple images, it creates smooth motion sequences while preserving realism, style, and fine details.
$0.300 / second
wan2.2-edit-video
Easily modify existing videos using simple text commands. With Wan 2.2 Video-Edit, you can change attire, character appearance, or other visual elements directly within your video—no need to start from scratch. Works on uploads of 480p or 720p, for up to two minutes.
$0.650 / second
wan2.6-image-to-video
WAN 2.6 Image-to-Video converts a single still image into a smooth, cinematic video clip. It preserves the original image’s composition, lighting, and style while adding natural motion, depth parallax, atmospheric effects, and gentle camera movement.
$0.200 / second
wan2.2-spicy-image-to-video
Wan2.2-spicy Image-to-Video transforms a single creative image into a short dynamic video with bold motion, stylized effects, high-contrast lighting, and energy-driven animations. The “spicy” variant produces more dramatic movement, more vivid colors, and more expressive visual effects.
$1.050 / second
happy-horse-1-video-edit-720p
Happy Horse 1.0 Video Edit (720p) - modify an input video at 720p using a natural-language instruction with optional reference images.
$0.100 / second
wan2.7-text-to-video
Alibaba WAN 2.7 Text-to-Video turns plain prompts into coherent, cinematic clips.
$0.100 / second
wan2.7-image-to-video
Alibaba WAN 2.7 converts images into videos with optional audio.
$0.100 / second
wan2.7-reference-to-video
Alibaba WAN 2.7 Reference-to-Video. Reference characters/props to generate new shots.
$0.100 / second
wan2.7-video-extend
Extend existing videos seamlessly with Wan 2.7.
$0.100 / second
wan2.7-video-edit
Perform prompt-driven video editing with multi-image reference support.
$1.800 / second
happy-horse-1-text-to-video-1080p
Happy Horse 1.0 Text to Video — generate expressive, stylized video clips from text prompts with vivid character motion and dynamic scene storytelling.
$1.800 / second
happy-horse-1-image-to-video-1080p
Happy Horse 1.0 Image to Video — bring still images to life with fluid, expressive animation and fine-grained motion control.
$0.900 / second
happy-horse-1-image-to-video-720p
Happy Horse 1.0 Image to Video (720p) — bring still images to life with fluid, expressive animation at 720p output resolution.
$0.900 / second
happy-horse-1-text-to-video-720p
Happy Horse 1.0 Text to Video (720p) — generate expressive, stylized video clips from text prompts at 720p output resolution.
$2.100 / second
happy-horse-1-reference-to-video-1080p
Happy Horse 1.0 Reference to Video (1080p) - generate expressive 1080p video clips conditioned on 1-9 reference images plus a text prompt.
$1.050 / second
happy-horse-1-reference-to-video-720p
Happy Horse 1.0 Reference to Video (720p) - generate expressive 720p video clips conditioned on 1-9 reference images plus a text prompt.
$0.900 / second
happy-horse-1.1-text-to-video-1080p
Happy Horse 1.1 Text to Video (1080p) — generate expressive 1080p video clips from text prompts with vivid character motion and dynamic scene storytelling.
$0.900 / second
happy-horse-1.1-image-to-video-1080p
Happy Horse 1.1 Image to Video (1080p) — bring still images to life with fluid, expressive 1080p animation.
$0.700 / second
happy-horse-1.1-image-to-video-720p
Happy Horse 1.1 Image to Video (720p) — bring still images to life with fluid, expressive 720p animation.
$0.700 / second
happy-horse-1.1-reference-to-video-720p
Happy Horse 1.1 Reference to Video (720p) — generate 720p video conditioned on 1-9 reference images plus a text prompt.
$0.900 / second
happy-horse-1.1-reference-to-video-1080p
Happy Horse 1.1 Reference to Video (1080p) — generate 1080p video conditioned on 1-9 reference images plus a text prompt.
$0.900 / second
happy-horse-1.1-video-edit-1080p
Happy Horse 1.1 Video Edit (1080p) — modify an input video using natural-language instructions with optional reference images.
$0.700 / second
happy-horse-1.1-video-edit-720p
Happy Horse 1.1 Video Edit (720p) — modify an input video using natural-language instructions with optional reference images.
$0.016 / second
wan2.2-5b-fast-t2v
Wan 2.2 Fast is a lightweight, high-speed version of the Wan 2.2 model, optimized for quick text-to-video generation. It trades some cinematic detail for rapid results, making it perfect for prototyping, previews, social media clips, and quick storytelling.
$0.200 / second
wan2.2-speech-to-video
WAN2.2 Speech-to-Video transforms a static image into a talking video by synchronizing lip movements and facial expressions with an audio input. Simply provide a character image along with a speech dialogue, and the model generates a natural, expressive video where the subject speaks your lines.
$0.300 / second
wan2.1-image-to-video
Animate static images into expressive video sequences with WAN 2.1. Upload any image and guide its transformation into a moving scene — great for bringing art, characters, or photos to life with smooth motion and consistent style.
$0.650 / second
wan2.5-image-to-video
WAN 2.5 Image-to-Video takes your image as the starting frame and turns it into a dynamic video, preserving realism, motion, and camera effects. Upload a static image, add a descriptive text prompt, and the model generates cinematic motion—camera pans, environmental movement, and realistic physics—across the result.
$0.650 / second
wan2.5-text-to-video
WAN 2.5 Text-to-Video transforms written prompts into cinematic video clips with dynamic motion, realistic physics, and natural animation. It can also generate characters delivering dialogue, making it ideal for storytelling, ads, and creative showcases.
$0.440 / second
wan2.5-image-to-video-fast
Convert a single static image into a cinematic short video with realistic motion, dynamic camera movement, and environmental effects. The Fast mode generates high-quality videos quickly, perfect for rapid prototyping, social media clips, and immersive visual storytelling from still images.
$0.440 / second
wan2.5-text-to-video-fast
Transform text prompts into short, cinematic videos with natural motion, realistic environments, and dynamic camera perspectives. Fast mode delivers quick, high-fidelity video generation, ideal for creative storytelling, concept visuals, and social media content.
$0.500 / second
wan3.0-image-to-video
Wan 3.0 Image to Video animates a source image with synchronized audio from a motion prompt, with an optional end-frame image, selectable resolution, aspect ratio, and duration.
$0.200 / second
wan2.2-spicy-video-extend
Wan-2.2-spicy Video Extend continues an existing video by generating new frames that match the original style but add stronger motion, bolder effects, and spicier dramatics.
$0.650 / second
wan2.6-text-to-video
WAN 2.6 Text-to-Video generates smooth, cinematic videos directly from text prompts. It’s designed for strong scene coherence, atmospheric depth, and fluid camera motion, making it ideal for fantasy and sci-fi worlds, surreal concepts, environmental storytelling, and dramatic visual sequences with rich lighting and motion.
$0.550 / second
wan3.0-spicy-image-to-video
Wan 3.0 Spicy Image to Video animates a source image into a bold, high-contrast, high-motion video with synchronized audio from a motion prompt, with an optional end-frame image, selectable resolution, aspect ratio, and duration.
$2.100 / second
happy-horse-1-video-edit-1080p
Happy Horse 1.0 Video Edit (1080p) - modify an input video at 1080p using a natural-language instruction with optional reference images.
$1.000 / second
wan2.6-image-to-video-spicy
Alibaba WAN 2.6 Spicy Image-to-Video animates a still image into a bold, high-motion clip with optional audio-guided generation, single- or multi-shot framing, and prompt-expansion.
$1.000 / second
wan2.7-image-to-video-spicy
Alibaba WAN 2.7 Spicy Image-to-Video animates a still image into a bold, high-motion clip with optional audio-guided generation, single- or multi-shot framing, and prompt-expansion.
$0.700 / second
wan3.0-prime-text-to-video
Wan 3.0 Prime Text to Video generates a higher-fidelity video with synchronized audio directly from a text prompt, with selectable resolution, aspect ratio, and duration.
$0.500 / second
wan3.0-reference-to-video
Wan 3.0 Reference to Video generates a video guided by a prompt plus up to 10 reference images, 5 reference videos, and 5 reference audios for visual and audio coherence.
$0.700 / second
wan3.0-prime-image-to-video
Wan 3.0 Prime Image to Video animates a source image with synchronized audio from a motion prompt at higher fidelity, with an optional end-frame image, selectable resolution, aspect ratio, and duration.
$0.550 / second
wan3.0-spicy-reference-to-video
Wan 3.0 Spicy Reference to Video generates a bold, high-contrast, high-motion video guided by a prompt plus up to 10 reference images, 5 reference videos, and 5 reference audios for visual and audio coherence.
Image Generation Models
$0.013 / 1K tokens
z-image-base
Z-Image Base is a general-purpose text-to-image model designed for reliable, high-quality image generation from natural language prompts. It focuses on clear composition, good prompt adherence, and versatile output across everyday scenes, product-style visuals, characters, and creative concepts.
$0.040 / generation
qwen-text-to-image-2512
Qwen Image Text-to-Image 2512 generates high-resolution, visually consistent images from text prompts. It focuses on strong scene structure, clean composition, and atmospheric lighting, making it well-suited for cinematic environments, surreal concepts, fantasy and sci-fi worlds.
$0.030 / 1K tokens
wan2.1-text-to-image
WAN 2.1 is a powerful AI model that transforms text prompts into high-resolution, photorealistic images. It excels at detailed object rendering, realistic lighting, and fine textures, making it ideal for visual content, concept art, advertising, and digital storytelling.
$0.050 / 1K tokens
wan2.7-text-to-image
Alibaba WAN 2.7 Text-to-Image generates high-quality images from text prompts with thinking mode for enhanced image quality.
$0.300 / generation
wan2.1-lora-i2v
Bring still images to life using WAN 2.1 LoRA I2V, which supports custom LoRA fine-tunes for identity consistency. Animate expressions, subtle movements, or full-body actions while preserving personalized features from the image and LoRA.
$0.300 / generation
wan2.1-lora-t2v
WAN 2.1 LoRA T2V enables users to generate videos from text prompts with custom-trained LoRA modules. Tailor the generation to specific characters, outfits, or animation styles — ideal for brand storytelling, fan content, and stylized animations.
$0.030 / 1K tokens
qwen-image
Generate high-quality, detailed images from text prompts in various styles — from realistic to artistic — perfect for creative visuals, product shots, and concept art.
$0.030 / generation
qwen-image-edit-plus
Qwen Image Edit Plus is an upgraded image-editing model that supports multiple image references and superior text editing. Powered by the 20B-parameter Qwen architecture, it allows changes like background swap, style transfer, object removal/addition, and precise text edits (bilingual: English/Chinese) while maintaining visual consistency and preserving details of the original images.
$2.500 / generation
z-image-lora-trainer
Train custom image LoRA models from your dataset with zip uploads, auto-tuned defaults, and fast iteration for brand, character, or IP looks.
$2.000 / generation
qwen-image-2512-lora-trainer
Train custom Qwen-Image-2512 LoRA models 10x faster with style, character, and object training. Upload a ZIP dataset with images to start.
$0.040 / 1K tokens
wan2.5-text-to-image
WAN 2.5 Text-to-Image generates high-quality, realistic or stylized images from textual descriptions. It supports detailed visual storytelling, cinematic compositions, and versatile styles — from portraits and product shots to landscapes and fantasy scenes.
$0.040 / generation
qwen-image-edit-2511
Qwen Image Edit 2511 performs precise, instruction-driven edits on an existing image while preserving composition, lighting, and overall style. It’s well-suited for object replacement, material changes, localized edits, and subtle scene adjustments with strong visual consistency and minimal artifacts.
$0.045 / generation
wan2.6-image-edit
WAN 2.6 Image Edit applies targeted, instruction-based edits to an existing image while preserving composition, perspective, and lighting. It’s ideal for object replacement, material changes, environment tweaks, and style adjustments with clean integration and minimal artifacts—keeping the original scene coherent and cinematic.
$0.007 / 1K tokens
z-image-turbo
Z-Image Turbo is a high-speed text-to-image model optimized for fast creative generation. It produces detailed, high-contrast, high-resolution images with strong stylization control. Ideal for rapid concept creation, visual exploration, product ideas, fantasy scenes, and cinematic composition tests. Designed for low latency and strong prompt adherence.
$0.004 / 1K tokens
z-image-p
Z-Image P is based on PiAPI's Qubico/z-image text-to-image model.
$0.050 / generation
wan2.7-image-edit
Alibaba WAN 2.7 Image Edit performs prompt-driven image editing with support for multiple-image references.
$0.040 / 1K tokens
qwen-image-2.0
Qwen 2.0 Text to Image model with enhanced realism.
$0.090 / 1K tokens
qwen-image-2.0-pro
Qwen 2.0 Pro Text to Image model with maximum realism and fidelity.
$0.090 / generation
qwen-image-2.0-pro-edit
Qwen 2.0 Pro Image Edit model with maximum precision and modifications.
$0.100 / 1K tokens
wan2.7-text-to-image-pro
Alibaba WAN 2.7 Text-to-Image Pro generates high-quality images up to 4K from text prompts with thinking mode for enhanced image quality.
$0.050 / generation
wan3.0-image-edit
Wan 3.0 Image Edit is an upcoming AI image-editing model. Confirmed API controls, output specifications, and pricing will be published at launch.
$0.030 / generation
qwen-image-edit
The Qwen Edit Image Model allows you to modify existing images using text-based editing prompts. Instead of generating from scratch, you can upload a base image and describe the desired changes (e.g., replacing objects, altering colors, adding new elements).
$0.050 / 1K tokens
wan3.0-text-to-image
Wan 3.0 Text to Image is an upcoming AI image model. Confirmed API controls, output specifications, and pricing will be published at launch.
$0.040 / generation
wan2.5-image-edit
The Wan2.5 Edit Image model allows you to transform existing images with precision and creativity. By providing an image along with an edit prompt, you can make realistic changes, enhancements, or stylistic adjustments—whether it’s altering objects, changing backgrounds, adding details, or applying an entirely new artistic style.
$0.040 / generation
qwen3-pro-image-to-image
Qwen 3.0 Pro Image to Image transforms and edits existing reference images according to text prompts with professional-grade precision.
$0.040 / 1K tokens
qwen3-pro-text-to-image
Qwen 3.0 Pro Text to Image generates high-fidelity professional photorealistic images and editorial artwork from prompts with enhanced detail.
$2.000 / generation
qwen-image-lora-trainer
Train custom Qwen-Image LoRA models 10x faster. Fine-tune styles, characters, or object concepts from a ZIP dataset with auto-tuned learning rates and steps.
$2.500 / generation
z-image-base-lora-trainer
Train custom Z-Image Base LoRA models from your dataset with zip uploads, auto-tuned defaults, and fast iteration for brand, character, or IP looks.
$0.040 / 1K tokens
wan2.6-text-to-image
WAN 2.6 Text-to-Image generates detailed, cinematic still images from text prompts. It focuses on strong composition, atmospheric lighting, and clear subject structure, making it suitable for fantasy and sci-fi environments, surreal concepts, architectural visuals, and dramatic world-building imagery.
$0.030 / 1K tokens
qwen3-text-to-image
Qwen 3.0 Text to Image generates high-fidelity photorealistic images and editorial artwork from prompts with intelligent prompt rewriting.
$0.030 / generation
qwen3-image-to-image
Qwen 3.0 Image to Image transforms and edits existing reference images according to text prompts with precise style and content preservation.
$0.040 / generation
qwen-image-2.0-edit
Qwen 2.0 Image Edit model with precise background modification and enhancements.
$0.100 / generation
wan2.7-image-edit-pro
Alibaba WAN 2.7 Image Edit Pro performs prompt-driven image editing with multi-image reference support and up to 2K output.
Other Utility Models
$0.020 / generation
qwen-image-text-to-image-lora
Qwen-Image Text-to-Image LoRA pairs the 20B MMDiT next-generation text-to-image model with custom LoRA adapters (up to 3 LoRAs) for fast style customization, refined aesthetics, and character consistency.
$0.040 / generation
qwen-image-edit-lora
Qwen Image Edit LoRA is a fine-tuned image editing model with custom LoRA support. Modifying or adding elements with fine-tuned styles.
$0.040 / generation
qwen-image-edit-plus-lora
Qwen-Image-Edit-Plus (2509) is 20B MMDiT image-to-image editor supporting multi-image edits, single-image consistency, and native ControlNet. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.
$0.010 / generation
z-image-base-text-to-image-lora
Z-Image-Base LoRA (6B) text-to-image and image-to-image generation model with external LoRA support (up to 3 LoRAs) and aspect ratio selection.
$0.010 / generation
z-image-turbo-text-to-image-lora
Z-Image-Turbo LoRA (6B) enables ultra-fast text-to-image generation with external LoRA support. Generate photorealistic images in sub-second latency while applying up to 3 LoRAs for custom styles.
$0.020 / generation
qwen-image-text-to-image-2512-lora
Qwen-Image Text-to-Image 2512 LoRA pairs the 20B MMDiT next-generation text-to-image model with custom LoRA adapters (up to 3 LoRAs) for fast style customization, refined aesthetics, and character consistency.
$0.010 / generation
z-image-turbo-image-to-image-lora
Z-Image-Turbo Image-to-Image LoRA transforms reference images with custom LoRA styles in sub-second time. Apply up to 3 LoRAs for personalized image transformation.
$0.040 / generation
qwen-image-edit-2511-lora
Qwen Image Edit 2511 LoRA is an enhanced version with custom LoRA support for personalized styles. It delivers stronger edit consistency, robust multi-person identity/pose consistency, custom LoRA styles, enhanced industrial/product design, and improved geometric reasoning for structure-preserving edits.