Models

The top AI models. One credit balance.

The agent picks the right model for each job. Or name the one you want.

VideoPremium

Seedance 2.5

For the one clip that has to be the best: hero product ads and close-up faces. Slower to render.

Speech and lip-syncFirst frameImage, video and audio references
4 to 30sLength, 15s on Creator
1080pMax quality
When to use it
VideoStandard

Seedance 2.0

The everyday model for real-looking UGC clips. A person saying real words, or a still brought to life.

Speech and lip-syncFirst frameImage, video and audio references
4 to 15sLength
1080pMax quality
When to use it
GPT Image 2.5 example: a podcast host at her desk with a microphoneAI-generatedMade with agent-mediaGPT Image 2.5
ImagePremium

GPT Image 2.5

For the image a video is built on: a portrait, a character sheet, a first frame, or an edit that keeps the face.

Text to imageImage editUp to 4 references
3Sizes
Same faceStrength
When to use it
GPT Image 2.5 Flare example: a man on his porch holds up a jar of honeyAI-generatedMade with agent-mediaGPT Image 2.5 Flare
ImageStandard

GPT Image 2.5 Flare

The same look as GPT Image 2.5, faster. For variants, draft frames and batches.

Text to imageImage editUp to 4 references
3Sizes
FastStrength
When to use it
Nano Banana Pro example: a speckled mug printed with the words agent media on a sunny deskAI-generatedMade with agent-mediaNano Banana Pro
ImagePremium

Nano Banana Pro

For edits, reference-based images and clean text inside the image.

Text to imageImage editUp to 4 references
1K, 2KSizes
Clean textStrength
GPT Image 2 example: a woman in her parked car holds up a plain travel mugAI-generatedMade with agent-mediaGPT Image 2
ImageStandard

GPT Image 2

The previous image model, kept for work that was built on it. New work uses GPT Image 2.5.

Text to imageImage editUp to 4 references
3Sizes
ContinuityStrength
When to use it
VoiceStandard

ElevenLabs

A clean voice track for narration and voiceover, when no face needs to speak on screen.

Text to speechMany languagesDelivery cues like [whispers]
Voice fileOutput
MultilingualStrength
When to use it

Not sure which one?

Just say what you sell. The agent picks the model for you.

Make my free video

Every example is a real output from our own account, checked October 18, 2026.