# 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.

- [Nano Banana 2.1](https://ai2face.com/models/nano-banana-2-1.md): Nano Banana 2.1 is Google DeepMind's image generation and editing model built on Gemini 3.6 Flash. Released on October 6, 2026, it replaces Nano Banana 2. It generates images up to 4K, supports up to 14 references, keeps up to 4 characters consistent, renders panoramas up to 8:1 and costs $0.0336 per 1K image in the API. Nano Banana 2.1 is available in the AI2Face generator.
- [Nano Banana 2](https://ai2face.com/models/nano-banana-2.md): Nano Banana 2 is Google DeepMind's model for generating and editing images, officially Gemini 3.1 Flash Image. Released on February 26, 2026, it offers near-Pro quality at Flash speed, resolutions from 512 px to 4K and up to 14 references. It's available in the AI2Face generator.
- [Nano Banana](https://ai2face.com/models/nano-banana.md): Nano Banana is Google's family of models for generating and editing images. The first one, Gemini 2.5 Flash Image, came out on August 26, 2025; today the family has four versions: Nano Banana, Nano Banana 2, 2 Lite and Pro. All four are available in the AI2Face generator.
- [Nano Banana 2 Lite](https://ai2face.com/models/nano-banana-2-lite.md): Nano Banana 2 Lite is Google's fastest and cheapest image model in the Nano Banana family, officially Gemini 3.1 Flash-Lite Image. Released on June 30, 2026, it makes a 1K image in about 4 seconds for $0.0336 in the API — half the price of Nano Banana 2. Resolution is 1K only. Nano Banana Lite is available in the AI2Face generator.
- [Nano Banana Pro](https://ai2face.com/models/nano-banana-pro.md): Nano Banana Pro is Google DeepMind's model for generating and editing images, officially Gemini 3 Pro Image. Released on November 20, 2025, it creates images up to 4K, renders text in images cleanly in many languages and accepts up to 14 references. In AI2Face it's available in the generator and the character builder.
- [Veo 3.1](https://ai2face.com/models/veo-3-1.md): Veo 3.1 is a video model from Google DeepMind, released on 15 October 2025. It makes clips of 4, 6 or 8 seconds at 720p, 1080p or 4K and generates sound with the picture: speech, effects and background. It can animate a photo, keep a character across three references and build a transition between a first and last frame. Veo 3.1, Fast and Lite are available in AI2Face.
- [Gemini Omni](https://ai2face.com/models/gemini-omni.md): Gemini Omni is a Google DeepMind video model introduced at Google I/O in May 2026. Omni 1.1 Flash (August 27, 2026) makes 3–10 second clips with sound from text, images and short videos, edits finished clips with ordinary phrases and extends them to 40 seconds. Google calls it “Nano Banana for video”. Omni Flash is available in the AI2Face generator.
- [Seedream](https://ai2face.com/models/seedream.md): Seedream is the family of image models from ByteDance Seed. Seedream 4.5 (December 2025) makes frames up to 4K and holds up to 10 references, 5.0 Lite (February 13, 2026) thinks before it generates and looks up information on the web, and 5.0 Pro (July 8, 2026) lays out complex designs, splits the picture into layers and writes text in 10+ languages. All three versions are in the AI2Face generator.
- [Seedance](https://ai2face.com/models/seedance.md): Seedance is a family of ByteDance Seed video models that generate a clip with sound from text, photos, video and audio. Seedance 2.0 came out on February 12, 2026 and makes 4–15 seconds. Seedance 2.5 came out on July 31, 2026: up to 30 seconds in one generation, two extensions and up to 50 references. Seedance is available in the AI2Face generator.
- [Kling](https://ai2face.com/models/kling.md): Kling is a family of video models from the Chinese company Kuaishou. Kling 3.0 launched on February 5, 2026: clips up to 15 seconds in 720p and 1080p, several shots in one generation, native audio in many languages, and characters from photos and video. On September 28, 2026, Kling announced version 4.0 with generation up to 30 seconds, still in early access. In AI2Face, Kling is available in the generator.
- [GPT Image](https://ai2face.com/models/gpt-image.md): GPT Image is the family of OpenAI image models behind picture generation in ChatGPT. GPT Image 2 launched on April 21, 2026, with accurate text in many languages, layout, infographics and comics. GPT Image 2.5 (September 8, 2026) comes in two versions, fast Flare and precise Sunburst, up to 50% faster and better at keeping faces from references. In AI2Face, GPT Image is available in the generator.
- [Midjourney](https://ai2face.com/models/midjourney.md): Midjourney is an image model from the independent lab of the same name, known for an artistic look. The current default is V8.2 as of July 24, 2026: a bolder aesthetic, personalization to the author's taste, HD up to 2048 px and a new Edit Model for edits from a reference. It also makes video, a clip from an image up to 21 seconds. In AI2Face, Midjourney is available in the generator.
- [FLUX](https://ai2face.com/models/flux.md): FLUX AI is a family of models from the German lab Black Forest Labs. FLUX.2 (November 25, 2025) generates and edits images up to 4 MP, takes up to 10 references and renders text and infographics, with open weights for the dev version. FLUX 3 Video (August 4, 2026) makes clips up to 20 seconds with sound and speech and ranks #3 on Arena's video leaderboard. FLUX is available in the AI2Face generator.
- [Hailuo](https://ai2face.com/models/hailuo.md): Hailuo AI is the video model from Chinese company MiniMax. Hailuo 2.3 (October 2025) makes 6–10 second clips in 768p and 1080p, while the new MiniMax H3 (July 31, 2026) makes 4–15 seconds up to 2K with stereo sound, up to 12 references and speech in 11 languages. H3's weights are open, and it ranks 8th of 48 on Arena. Hailuo is available in the AI2Face generator.
- [Wan](https://ai2face.com/models/wan.md): Wan AI is Alibaba's family of video models. Wan 3.0, officially launched on August 24, 2026, makes clips up to 30 seconds with sound — speech, effects and music — and takes up to 20 references: photos, videos, audio and even documents. It ranks 6th of 48 on Arena. The earlier Wan 2.1 and 2.2 are open under Apache 2.0. Wan is available in the AI2Face generator.
- [Runway](https://ai2face.com/models/runway.md): Runway AI is the family of video models from US company Runway. Gen-4.5 launched on December 1, 2025, and topped the Artificial Analysis leaderboard at release: 2–10 second clips in 720p from text and images, complex camera choreography and precise prompt adherence. Aleph 2.0 edits finished videos, and Gen-3 was retired in July 2026. Runway is available in the AI2Face generator.
- [Grok Imagine](https://ai2face.com/models/grok-imagine.md): Grok Imagine is xAI's image and video generator. Imagine Image 2.0 came out on August 7, 2026: precise typography, region editing, background removal and up to 5 references, ranking 6th of 80 on Arena. Imagine Video 1.5 makes 1080p clips with sound, and since July it supports face and voice references. Grok Imagine is available in the AI2Face generator.