Skip to content

AI models for image and video generation

We break down popular image and video generation models: what they can do, how they differ and which tasks they suit. Every model page has key facts, prompt examples and answers to common questions.

A stack of printed photos on a dark desk — a portrait, a product shot, a mountain panorama — with a ripe banana on top

Google DeepMindImagesAvailable in AI2Face

Nano Banana 2.1

Google's default image model — up to 4K, 14 references, 8:1 panoramas, half the price of Nano Banana 2.

Aerial shot of a misty green valley with a winding river and clouds on the slopes

Google DeepMindImagesAvailable in AI2Face

Nano Banana 2

Google's main photo model — Flash speed, near-Pro quality, up to 4K.

Photos, a watercolor and a clay fox figurine on a wooden table next to a banana

Google DeepMindImagesAvailable in AI2Face

Nano Banana

Google's family of models for generating and editing photos — from the fast Nano Banana to the flagship Pro.

A white desk from above: a sketchbook with mountain landscape thumbnails, a smartphone with a color landscape, a stopwatch, markers and an espresso

Google DeepMindImagesAvailable in AI2Face

Nano Banana 2 Lite

Google's lightweight image model — results in seconds, 1K only, the cheapest in the family.

A designer in a studio at night looks at a grid of generated portraits and posters on a monitor

Google DeepMindImagesAvailable in AI2Face

Nano Banana Pro

Google's flagship model for photorealism, accurate text in images and work with references in 4K.

A white lighthouse on a cliff in a storm, waves breaking on the rocks, a beam of light through the rain

Google DeepMindVideoAvailable in AI2Face

Veo 3.1

A Google video model with sound — speech, effects and music in the same clip, up to 4K and 8 seconds.

Building a Roman aqueduct: workers lift stone blocks with a treadwheel crane, arches on wooden centering

Google DeepMindVideoAvailable in AI2Face

Gemini Omni

“Nano Banana for video” — a clip with sound from text, photos and video, and edits in ordinary phrases.

A design studio desk in the evening: posters, fabric swatches, a vase of dried flowers and a brass lamp by a window with city lights

ByteDance SeedImagesAvailable in AI2Face

Seedream

ByteDance image models — photorealism, design and text in 10+ languages, up to 4K.

An elderly fisherman in a yellow raincoat and a sea cap on a boat in a foggy harbor at dawn

ByteDance SeedVideoAvailable in AI2Face

Seedance

ByteDance's video model with sound — up to 30 seconds per generation and up to 50 references in 2.5.

A chef tosses pasta in a pan with flames in a trattoria kitchen

KuaishouVideoAvailable in AI2Face

Kling

Kuaishou's video model: multi-shot clips up to 15 seconds with sound. Kling 4.0 promises 30 seconds.

Top view of a creative agency desk: poster sketches, a tablet with a drawing, markers, stickers and coffee

OpenAIImagesAvailable in AI2Face

GPT Image

OpenAI's image model in ChatGPT — accurate text, layout and edits, up to 3840 px.

Surreal landscape: a giant koi made of clouds swims above a misty mountain lake, a boat with a lantern below

MidjourneyImagesAvailable in AI2Face

Midjourney

An artistic image model: a recognizable look, styles, personalization and HD at 2K.

A sunny artist's loft: a large table with photo prints and a palette, a laptop showing a mountain landscape, plants by the window

Black Forest LabsImagesAvailable in AI2Face

FLUX

BFL's models — FLUX.2 for 4 MP images with 10 references, FLUX 3 Video ranked #3 for video.

A red convertible drives along a winding road above a turquoise sea at sunset, aerial view

MiniMaxVideoAvailable in AI2Face

Hailuo

MiniMax's video models — H3 makes 15-second 2K clips with stereo sound, with open weights.

A young woman in a green scarf steps off a red train at a snowy mountain station

AlibabaVideoAvailable in AI2Face

Wan

Alibaba's video model — clips up to 30 seconds with sound, up to 20 references, even documents.

A dark editing suite: a monitor shows a rider in the desert at sunset, next to a control panel with glowing buttons and headphones

RunwayVideoAvailable in AI2Face

Runway

Runway's video models — Gen-4.5 for text- and image-to-video with precise camera choreography.

A skateboarder does a kickflip over a concrete staircase in a city plaza at sunset

xAIImagesAvailable in AI2Face

Grok Imagine

Images and video from xAI — Image 2.0 ranks #6 on Arena, Video 1.5 adds sound and references.