Models

Balanced speed and quality. Text-to-video and image-to-video with optional reference images and voice narration.

Lower-cost relaxed-priority variant of Veo 3.1 Fast.

Highest-fidelity Veo 3.1 tier for production-grade output.

Fastest and most affordable Veo variant. 4–8 second clips.

Lowest-cost relaxed-priority Veo tier.

Continues an existing clip by up to 8 more seconds. Requires a completed task_id.

Upscales a completed video to 1080p or 4K.

Text-to-video and image-to-video with 768p/1080p output.

Latest Hailuo generation with improved motion and detail.

Speed-optimized Hailuo variant. Image-to-video focused.

Multimodal video with elements, start/end frame, transform and video-reference modes.

Edit an existing video with reference or transform mode.

Reasoning-enhanced video generation. Up to 10 seconds.

Edit an existing video with reference or transform mode.

Flagship video model. Up to 15s, multi-shot, native audio.

Image-to-video with voice support and native audio.

Fast text-to-video and image-to-video with sound effects.

Image-to-video specialist. Animates a start frame into 5 or 10 seconds.

Pro-only premium image-to-video.

Transfer motion from a reference video onto a character image.

High-fidelity motion transfer with facial consistency.

Extend a previously generated video via preset or custom mode.

AI video upscaling with selectable models and frame interpolation.

Fast, affordable image generation and editing with strong prompt adherence.

Higher-fidelity image generation with 1K/2K/4K resolution.

Renders text in 50+ languages. Resolution-based (1K/2K/4K) with reference images.

High-quality image generation with up to 8 reference images.

Flexible generation with steps and cfg controls.

Maximum quality image generation.

Reasoning-enhanced image generation.

High-quality image generation with reference support.

Batch image generation.

Image generation with restyle, portrait, character and elements modes.

AI image upscaling with selectable presets and face recovery.

Prompt-guided generative image upscaling.

Best quality with text rendering and realistic textures.

Faster variant of 2.0 Pro with the same capabilities.

Highest realism with fewest AI artifacts.

Diverse artistic styles with fast generation.

Base Qwen image model with text-to-image and editing.

Lightweight fast text-to-image with Chinese/English text rendering.

Highest quality with thinking mode and up to 4K.

Faster variant of 2.7 Pro, max 2K resolution.

Image generation with style transfer and editing.

Text-to-image and editing.

Text-to-image generation.

Fast and affordable text-to-image.

Speed-focused image generation and editing.

Generate full songs from a prompt with lyrics and style.

Add vocals to an uploaded instrumental track.

Add instrumental backing to uploaded vocals.

Extend an existing song with a new section.

Create a cover version of an existing song.

Extract vocals and instrumental stems from a track.

Extract all individual stems from a track.

Generate lyrics from a prompt.

Export a clip as WAV, MP3 or M4A.

Neural text-to-speech with a large voice library.

Multi-voice dialogue generation.

Generate sound effects from a text prompt.

Isolate vocals from an audio track.

Transcribe audio to text.

Coding model with strong agentic tool use.

Faster, lower-cost coding model.

Fast, affordable general-purpose model.

Smallest, fastest GPT 5.4 tier.

Flagship general-purpose language model.

Next-generation flagship language model.

Balanced GPT 5.6 tier.

High-capability GPT 5.6 tier.

Top-tier GPT 5.6 model.

Creative writing and storytelling model.

Most capable Claude model for complex reasoning.

Fast, lightweight Claude model.

High-capability reasoning model.

Balanced Claude model.

Next-generation balanced Claude model.

Strong coding and reasoning model.

High-capability reasoning model.

High-capability reasoning model.

Latest high-capability Claude model.

Long-context reasoning model (1M context).