AOKIstudio

Comparison of AI Image and Video Generators

A real economic ecosystem is developing around generative images and videos, particularly thanks to open-source models like Stable Diffusion and online hosting solutions, which allow everyone to create, with minimal investment, websites and services for generating AI images and videos. The goal of this comparison is not to encourage the use of AI generators, but to provide an objective perspective on the multitude of available AI tools. Which platforms own their own models? Which ones integrate open-source models? Which platforms are transparent about their mode of operation? Why are some platforms completely opaque? Discover and compare the specifics of the main AI image and video generators.
In order to objectively compare AI image generation programs, all presentation images were generated from a single prompt, the one used on OpenAI's website to compare DALL·E 2 and DALL·E 3: "An expressive oil painting of a basketball player dunking, depicted as an explosion of a nebula". Common features among AI image and video generators are subscription pricing, based on the number of images or videos generated, as well as the ability to copy-paste or modify prompts from other users to generate your own content.

NightCafe

Text → Image|Predefined AI Styles|Developed by NightCafe, Australia, since November 2019

AI Model Used: Stable Diffusion, DALL·E, Ideogram, Imagen, Flux.

Images generated by NightCafe, January 2025, CC0

Overshadowed by MidJourney and DALL·E, NightCafe is one of the first consumer-facing AI image generators that does not require local installation, unlike experimental AI models like DeepDream (Google, 2015) or StyleGAN (NVIDIA, 2018).
Continue reading...

Dream

Text → Image|Predefined AI Styles|Developed by Wombo, Canada, since February 2021

AI Model Used: VQGAN then Stable Diffusion (Developed by Stability AI)

Images generated by Dream (style Cartoon V3), January 2025, CC0

Dream, originally using the VQGAN (Vector Quantized Generative Adversarial Network) AI model, is one of the first tools for generating images from text descriptions. Usable today via a web browser or mobile app (iOS and Android), Dream offers over 150 predefined AI styles, such as cartoon, surrealist, impressionist, abstract, baroque, horror, bad trip, isometric.
Continue reading...

Craiyon

Text → Image|Developed by Craiyon, United States, since July 2021

AI Model Used: Craiyon (Proprietary model developed by Craiyon)

Images generated by Craiyon, January 2025, CC0

Craiyon was originally an open-source project developed for a competition organized by Hugging Face in July 2021. To avoid confusion with the DALL·E model, which inspired it, its original name, DALL·E mini, was rebranded as Craiyon in July 2022.
Continue reading...

Midjourney

Text → Image|Developed by Midjourney, United States, since June 2022

AI Model Used: MidJourney (Proprietary model developed by MidJourney)

Images generated by MidJourney, January 2025, CC0

MidJourney is one of the top three consumer-facing AI image generators alongside DALL·E and Stable Diffusion. Some images generated by MidJourney, such as Emmanuel Macron picking up trash, Donald Trump being arrested by police, or the Pope wearing a puffer jacket, have gone viral worldwide.
Continue reading...

DALL·E

Text → Image|Developed by OpenAI, United States, since April 2022 (DALL·E 2)

AI Model Used: DALL·E 3 (Proprietary model developed by OpenAI)

Images generated by DALL·E 3, January 2025, CC0

DALL·E, an image generator developed by OpenAI, stands out from other generators due to its integration with ChatGPT. Users can generate images from text descriptions (prompts) and then modify them through conversation, directly within the ChatGPT interface.
Continue reading...

DreamStudio

Text → Image|Image → Image|Developed by Stability AI, United Kingdom, since August 2022

AI Model Used: Stable Diffusion XL (Developed by Stability AI)

Images generated by DreamStudio, January 2025, CC0

DreamStudio is the online platform of Stability AI, primarily known for developing Stable Diffusion, and allows users to generate images from text descriptions using its open-source generative AI model, Stable Diffusion.
Continue reading...

Neural Love

Text → Image|Image → Video|Developed by Neural Love, Poland, since September 2022

AI Model Used: Stable Diffusion

Images generated by Neural Love, January 2025, CC0

Since 2020, Neural Love has been offering AI-powered online services for enhancing and restoring photos and videos, and has provided an AI image generator since September 2022.
Continue reading...

Runway

Text → Image|Text → Video|Image → Video|Video → Video|Developed by Runway, United States, since October 2022

AI Model Used: Gen (Proprietary model developed by Runway)

Images generated by Runway ML, January 2025, CC0

Originally, in 2018, Runway ML (Runway Machine Learning) was not an image generator but offered AI tools integrated into software like After Effects, particularly for background removal and style transfer. Runway AI (Artificial Intelligence) now offers a suite of AI tools, including content generation and editing, as well as the ability to share workspaces and collaborate with other users on the same project. Runway also hosts an annual AI film festival and provides Runway Watch, an online platform dedicated to showcasing AI-created content.
Continue reading...

OpenArt

Text → Image|Image → Video|Developed by OpenArt, United States, since October 2022

AI Model Used: Stable Diffusion

Images generated by OpenArt, January 2025, CC0

Founded by former Google employees, OpenArt, a platform dedicated to generative art, offers a wide range of AI-based tools. For example, it is possible to upload reference images to the site and guide the generation of images toward specific poses or a particular style.
Continue reading...

Recraft

Text → Image|Developed by Recraft, United Kingdom, since October 2022

AI Model Used: Information not available

Images generated by Recraft, January 2025, CC0

Recraft, with over 3 million users and 350 million images generated by early 2025, aims to become "an indispensable daily tool for every designer."
Continue reading...

Dreamup

Text → Image|Developed by DeviantArt, United States, since November 2022

AI Model Used: Stable Diffusion

Images generated by Dreamup, January 2025, CC0

DreamUp is developed by DeviantArt, a website dedicated to digital art creation, founded in 2000, and home to a community of nearly 100 million registered members, allowing everyone to showcase and share their personal works and creations.
Continue reading...

Fotor

Text → Image|Developed by Fotor, United States, since November 2022

AI Model Used: Information not available

Images generated by Fotor, January 2025, CC0

Fotor, launched in 2012 as an image editing software, has become a multifunctional platform following the addition, in November 2022, of an AI image generator capable of creating visuals from text. It now offers numerous AI-based features.
Continue reading...

Getimg

Text → Image|Text → Video|Developed by Getimg, Unknown country, since November 2022

AI Model Used: Flux and Stable Diffusion XL

Images generated by Getimg, January 2025, CC0

Getimg is completely opaque about the people behind the project and the country of origin.
This lack of transparency leads us not to provide a dedicated presentation page for this AI tool.

Deepai

Text → Image|AI Detector|Text → Voice Synthesis (Voice)|Developed by Deepai, United States, since December 2022

AI Model Used: Information not available

Images generated by Deepai, January 2025, CC0

DeepAI, a small structure founded in 2016 by Kevin Baragona, has been offering tools and services for developers and AI enthusiasts since 2016 and, since December 2022, an AI image generator.
Continue reading...

Leonardo

Text → Image|Image → Video|Predefined AI Styles|Developed by Leonardo, Australia, since December 2022

AI Model Used: Phoenix (Proprietary model developed by Leonardo), Flux and Stable Diffusion (Optimized by Leonardo AI)

Images generated by Leonardo, January 2025, CC0

Leonardo, a generative AI platform, offers numerous tools such as inpainting, unzoom, image upscaling, background removal, and masking. In 2025, the Flow State tool provides a real-time, continuous, and infinite stream of images from a single prompt.
Continue reading...

Firefly

Text → Image|Text → Video|Text → Audio|Developed by Adobe, United States, since March 2023

AI Model Used: Adobe Sensei (Proprietary model developed by Adobe)

Images generated by Firefly in Adobe Illustrator, January 2025, CC0

Firefly, integrated into Adobe software (Photoshop, Illustrator, Premiere Pro...), allows users to generate and modify images, videos, audio files, and special effects using AI. Adobe places particular importance on copyright, and Firefly is trained exclusively on licensed data and Adobe Stock, unlike other AIs like MidJourney or Stable Diffusion, which exploit web images normally protected by copyright.
Continue reading...

PromeAI

Drawings → Realistic image|Developed by LibAI, Hong Kong, since April 2023

AI Model Used: Information Not Available

Images generated by PromeAI, January 2025, CC0

LibAI, founded in 2019, specializes in the architectural design and creative sectors. Since April 2023, it has offered a service that transforms drawings and sketches into realistic photos.
Continue reading...

Imagine

Text ↔ Image|Text → Image|Developed by Imagine since May 2023

AI Model Used: Information not available

Images generated by Imagine, January 2025, CC0

Imagine is one of the AI image and video generators strangely opaque about its origin and operation. Perhaps it prefers to remain discreet about its location to avoid potential legal restrictions or controversies related to the origin of its models. The information "May 2023" comes from a search on Wayback Machine (Archive.org).
This lack of transparency leads us not to provide a dedicated presentation page for this AI tool.

123RF

Text → Image|Developed by Inmagine, Singapore, since May 2023

AI Model Used: Information not available

Images generated by 123RF, January 2025, CC0

Founded in 2005, 123RF, with over 200 million royalty-free images and photographs, is one of the leaders in digital stock agencies. Since 2023, 123RF has offered a range of AI-powered tools, such as AI image generation and variation, or AI background removal.
Continue reading...

Ideogram

AI Image Generator|Developed by Ideogram, Canada, since August 2023

AI Model Used: Ideogram AI (Proprietary model developed by Ideogram)

Images generated by Ideogram, January 2025, CC0

Ideogram's Magic Prompt tool automatically optimizes initial text instructions by adding contextual details, atmospheres, and stylistic information, and Batch Generation allows for the production of large quantities of images without manual intervention, with the import of a spreadsheet file containing all the prompts.
Continue reading...

Krea

Text → Image|Text → Video|Image → Upscale|Developed by Krea, Spain, since November 2023

AI Model Used: Stable Diffusion

Images generated by Krea, January 2025, CC0

Krea offers a range of features, including real-time image generation, image enhancement and scaling, as well as video generation. It also allows users to train custom AI models based on their styles.
Continue reading...

Pixlr

Text → Image|Developed by Pixlr, Singapore, since December 2023

AI Model Used: Information not available

Images generated by Pixlr, January 2025, CC0

Pixlr, an online image editing tool launched in 2008, has offered AI tools since December 2023, such as text-to-image generation, generative fill, background removal, and image resizing.
Continue reading...

Freepik

Text → Image|Text → Video|Text → Voice Synthesis (Voice)|Developed by Freepik, Spain, since December 2023

AI Model Used: Flux

Images generated by Freepik, January 2025, CC0

Freepik is a search engine for graphic resources and an image bank founded in 2010. Its AI tool, Pikaso, allows users to generate images, videos, and voiceovers. Interestingly, it is one of the few AI programs to generate images of female basketball players and not just men.
Continue reading...

ImageFX

Text → Image|Developed by Google DeepMind, United States, since February 2024

AI Model Used: Imagen (Proprietary model developed by Google DeepMind)
Continue reading...

Images generated by ImageFX, January 2025, CC0

ImageFX, signed by Google, like TextFX, Music-FX, Video-FX, Whisk, and many other projects, is part of Google Labs, an experimental space for Google's latest AI technologies.
One of the unique features of ImageFX is marking generated images with SynthID, a digital watermark that allows for the traceability of AI content without altering its appearance.

Akool

Text → Image|Developed by Akool, United States, since April 2024

AI Model Used: Information not available

Images generated by Akool, January 2025, CC0

AKOOL, a generative AI platform specializing in the creation of visual content for marketing and advertising, has allowed users to generate AI images since April 2024.
Continue reading...

KlingAI

Text ↔ Image|Text → Video|Developed by Kuaishou Technologys, China, since May 2024

AI Model Used: Information not available

Images generated by KlingAI, January 2025, CC0

KlingAI is an AI platform developed by Kuaishou Technology, a Chinese company specializing in short video sharing.
Continue reading...

Wirestock

Text → Image|Developed by Wirestock, United States, since May 2024

AI Model Used: Google's Imagen

Images generated by Wirestock, January 2025, CC0

Founded in 2018, Wirestock is an online platform that simplifies the process of monetizing visual content for creators. By eliminating traditional manual steps, such as registering on multiple marketplaces, SEO, and payment management, Wirestock allows artists to more easily sell their photos, videos, illustrations, and AI-generated content.
Continue reading...

Vivago

Text ↔ Image|Text ↔ Video|Image → Video|Text → 3D|Developed by Sparking Innovations, Hong Kong, since May 2024

AI Model Used: Information not available

Images generated by Vivago, January 2025, CC0

Vivago offers classic AI tools, such as removing unwanted parts of an image, expanding an image, improving its resolution, or removing unwanted parts, but also new tools, although still unconvincing and not usable in professional contexts, such as generating 3D models from text queries.
Continue reading...

Artbreeder

Text → Image|Developed by Morphogen, United States, since May 2024

AI Model Used: BigGAN and StyleGAN

Images generated by Artbreeder, January 2025, CC0

Artbreeder is an example of an AI image generation platform developed collaboratively by a small team: the Morphogen collective, founded by Joël Simon after his studies in computer science and fine arts.
Artbreeder has offered a text-to-image generation tool since May 2024, but already allowed users to mix two images together as early as August 2019.
Continue reading...

Dream Machine

Text → Image|Text → Video|Developed by Luma AI, United States, since June 2024

AI Model Used: Information not available

Images generated by Dream Machine, January 2025, CC0

Luma AI, founded in 2021, launched its AI image and video generator, Dream Machine, in June 2024.
Continue reading...

Designer

Text ↔ Image|Image → Image|Developed by Microsoft, United States, since July 2024

AI Model Used: DALL·E

Images generated by Canva, January 2025, CC0

Following its significant investments in OpenAI, Microsoft uses ChatGPT for its Bing search engine and DALL·E for Microsoft Designer. Note the similarity between images generated by ChatGPT and Microsoft Designer, confirming the use of DALL·E.
Continue reading...

Canva

Text ↔ Image|Developed by Canva, Australia, July 2024

AI Model Used: Leonardo

Images generated by Canva, January 2025, CC0

Canva, specializing in online graphic design since 2010, integrated the Leonardo model in July 2024 to enhance its features and offer AI image generation.
Continue reading...

Envato

Text ↔ Image|Image → Upscale|Image → Erase|Developed by Envato, Australia, since September 2024

AI Model Used: AI ImageGen (Proprietary model developed by Envato)

Images generated by Envato, January 2025, CC0

Envato, a marketplace for digital creative resources and templates founded in 2006, launched its own AI-based image generator in 2024.
Continue reading...

Flux

Text ↔ Image|Image ↔ Prompt|Text → Video|Image → Upscale|Developed by Black Forest Labs, Germany, since August 2024

AI Model Used: Flux AI (Proprietary model developed by Black Forest Labs)

Images generated by Flux, January 2025, CC0

Founded by former Stable Diffusion employees, Flux, an open-source AI, allows users to generate images and videos from text prompts.
Continue reading...

Domo AI

Text → Image|Image → Video|Video → Animation|Developed by Domo AI, Singapore, since September 2024

AI Model Used: Information not available

Images generated by Domo AI, January 2025, CC0

Since December 2023, Domo AI has been offering tools for modifying videos into stylized animations, such as cartoon, anime, 3D... and, since September 2024, an AI-powered text-to-image generator.
Continue reading...

Starryai

Text → Image|Upscale|Developed by Starryai, unknown country, unknown creation date

AI Model Used: VQGAN+CLIP and CLIP-Guided Diffusion

Images generated by Starryai, January 2025, CC0

Starryai is completely opaque regarding the people behind the project, the country of origin, and the creation date.
This lack of transparency leads us to not offer a dedicated presentation page for this AI tool.

Janus

Text → Image|Developed by DeepSeek, China, since January 2025

AI Model Used: Janus-Pro (Proprietary model developed by DeepSeek)

Images generated by Janus, January 2025, CC0

Janus, available as open source, is an advanced multimodal model designed to handle both image understanding and text-to-image generation.
Continue reading...

Artlist

Text → Image|Text → Video|Developed by Artlist, Israel, since April 2025

AI model used: Information not available

Images generated by Artlist, April 2025, CC0

Artlist has been offering royalty-free digital content since 2016, including original music and sound effects. The company launched an AI-powered voiceover generator in June 2024, the AI Assistant extension for Adobe Premiere Pro in November 2024, and an AI image and video generator in April 2025.
Read more...

Other Platforms

It is difficult to assess the number of AI image generation programs currently available, and we have chosen not to write articles about platforms that are deliberately opaque regarding the identity of their creators and their operating methods.
Among these many opaque sites: Designs.ai, Lexica.art, Imgfx.ai, Getimg.ai, Imagine.art, Imagineme.ai, Artistly.ai, Artguru.ai...

Platforms Under Review

Hitpaw (AI video upscaler)
Pollo (Text → Video)
Sora (Text → Video)
Wonder Dynamics (Video → Animation)
Domoai (Video → Animation)...

Getty Images, Shutterstock and iStock (no free trial available for image generation) also offer image generation and editing tools (NVIDIA AI model).