Stability AI

Stability AI is an artificial intelligence company founded in 2019 by Emad Mostaque, known for its work in generative AI, particularly with its text-to-image model, Stable Diffusion. The company aims to build open AI tools that allow people to reach their full potential by combining collective intelligence and augmented technology. Stability AI operates in the field of research services and is headquartered in London, England.

The company gained significant recognition in August 2022 with the release of Stable Diffusion, an AI model that generates images from text descriptions. This model is widely used for creative applications and is available under a community license. In March 2024, Emad Mostaque stepped down as CEO, with Shan Shan Wong (COO) and Christian Laforte (CTO) being appointed as interim co-CEOs. James Cameron, the Academy Award-winning filmmaker, also joined the Board of Directors, adding to the company's credibility in the creative sector.

Stability AI has raised $80 million in funding, with 11 total investors, including major players like Coatue, Lightspeed Venture Partners, and O’Shaughnessy Ventures LLC. 

The company offers a wide range of AI models, including tools for image, video, audio, 3D, and text generation. Apart from Stable Diffusion, Stability AI's product suite includes Stable Audio 2.0 for audio generation, Stable Video 4D for multi-angle video creation, and Stable LM for language processing.

The company provides APIs for image services like object erasure, background removal, and inpainting. Stability AI offers various pricing tiers, including free community licenses for non-commercial use and enterprise licenses for larger organisations.

Stability AI has faced legal challenges, notably from Getty Images, which accused the company of using over 12 million of its photos without proper licensing to train Stable Diffusion. This lawsuit is part of a broader concern about the use of copyrighted material in AI training.

The company has formed partnerships with technology giants like AWS, Google, Intel, and NVIDIA to scale and optimise its AI models. Its foundation models are available on Amazon’s Bedrock and SageMaker JumpStart. Stability AI continues to expand its research efforts, publishing papers on advancements in AI, including its Hourglass Diffusion Transformer and Adversarial Diffusion Distillation.

In 2024, Stability AI announced significant new investments, which helped fuel further developments in its model suite. Key updates included the release of Stable Diffusion 3. Additionally, Stable Fast 3D was introduced, allowing rapid 3D asset creation from single images. 

Stability AI was founded in 2019 by Emad Mostaque to build open AI tools that can unlock human potential. The company focuses on creating artificial intelligence solutions, particularly in generative AI, using collective intelligence and augmented technology. It started with internal funding from Mostaque and later gained wider attention in August 2022 with the launch of its Stable Diffusion model, a text-to-image AI tool that allows users to generate images from textual descriptions. Stable Diffusion quickly became one of the most popular AI models in the field of generative AI.

In its early stages, Stability AI was supported largely by Mostaque's investments. However, a major turning point came when the company secured a $101 million investment led by Coatue, Lightspeed Venture Partners, and O'Shaughnessy Ventures LLC. This influx of funding allowed the company to scale its operations and enhance its AI models across various domains, including text, video, audio, and 3D generation.

One of Stability AI's most notable products, Stable Diffusion, has been applied across industries, from media and entertainment to architecture and retail. The company also introduced several other models such as Stable Audio for generating music and sound, Stable Video 4D for creating multi-angle videos, and Stable Fast 3D, a model that can produce 3D assets from single images. These models have been widely adopted by creators, developers, and enterprises globally.

In March 2024, Emad Mostaque resigned as CEO, and the company appointed Shan Shan Wong (COO) and Christian Laforte (CTO) as interim co-CEOs. Around the same time, Stability AI expanded its board of directors by including notable figures such as James Cameron, a renowned filmmaker, and Sean Parker, the Executive Chairman of Stability AI.

Stability AI has faced some legal challenges, most notably from Getty Images, which filed a lawsuit claiming the company had used over 12 million of its photos to train Stable Diffusion without proper licensing. Despite this, Stability AI continued its work and growth, with its models gaining more traction across different sectors.

In 2024, the company released Stable Diffusion 3 and continued to expand its model suite with Stable Audio 2.0 and Stable Video 4D, among others. Stability AI has formed partnerships with major tech companies like AWS, Google, Intel, and NVIDIA to help optimise and scale its AI models. Its models are available on platforms such as Amazon Bedrock and SageMaker JumpStart.

Today, Stability AI continues to develop open-source generative AI models, focusing on making these technologies accessible to everyone. It remains a leader in the field, with a strong emphasis on open innovation and the broad applicability of its models across different industries.

Stability AI's mission is to build open, accessible artificial intelligence tools that allow individuals and organisations to unlock their full potential. By developing state-of-the-art AI models across various fields, including image, video, audio, text, and 3D generation, the company seeks to democratise creativity and innovation. Stability AI aims to make these tools available to everyone, promoting transparency, collaboration, and collective intelligence. Through their open models, they strive to empower developers, researchers, and creators worldwide, enabling them to create, innovate, and solve real-world problems using cutting-edge AI technologies.

Stability AI's vision is to create a world where artificial intelligence is open, accessible, and beneficial for all. The company aspires to be a leader in the AI field by providing models that drive innovation and creativity across industries. Their goal is to ensure AI tools are available to everyone, regardless of technical expertise, enabling people to harness the power of AI in various sectors like entertainment, education, research, and business. Stability AI envisions a future where AI is used responsibly and ethically to improve human potential and solve global challenges.

Stability AI offers a wide range of artificial intelligence products and services, focusing on generative AI models that span across image, video, audio, text, and 3D generation. These models are designed to help individuals, businesses, and industries create content more easily, efficiently, and with greater creative potential. Below is an overview of their key products and services:

Stable Diffusion
Stable Diffusion is Stability AI’s flagship product, a text-to-image model that allows users to generate images from text prompts. Released in 2022, it has become a popular tool for creators and developers looking to produce photorealistic or artistic images based on their input. The model is available in various versions, including Stable Diffusion 3, which offers enhanced performance in generating images, especially when handling complex multi-subject prompts. The model’s small size also makes it compatible with consumer-grade PCs and enterprise-level GPUs. Stable Diffusion is available under a community license, making it accessible for non-commercial use, with commercial licenses offered for businesses.

Stable Audio 2.0
Stable Audio 2.0 is an advanced audio generation model that enables the creation of music, sound effects, and soundscapes through text prompts. This tool allows users to generate up to three minutes of high-quality audio in 44.1 kHz stereo. Stability AI has also introduced features like audio-to-audio generation, where users can upload audio samples and transform them using natural language instructions. This model is particularly useful for musicians, sound designers, and creators working in media production, enabling them to generate original audio content or enhance existing sounds.

Stable Video 4D
Stable Video 4D is a video generation model that allows users to create dynamic, multi-angle videos from a single video input. This model generates multiple new views from different camera angles, making it a powerful tool for content creators in film, advertising, and virtual reality. The model is capable of producing high-quality videos in just 40 seconds, enhancing creativity in video production. Stable Video 4D supports both text-to-video and image-to-video workflows, allowing users to turn concepts into cinematic experiences.

Stable Fast 3D
Stable Fast 3D is a model designed to create high-quality 3D assets from single images. It is particularly useful for game developers, architects, and professionals in the retail and design industries who need to create detailed 3D objects quickly. Stable Fast 3D can produce 3D models in 0.5 seconds, providing a fast and efficient way to visualise and interact with 3D objects. This model is also used in virtual reality (VR) and augmented reality (AR) applications, expanding its utility across multiple industries.

Stable LM (Language Models)
Stability AI has developed several language models under the Stable LM series, including Stable LM 2 12B and Stable LM 2 1.6B. These models are trained on multilingual data, supporting languages such as English, Spanish, German, French, and more. The Stable LM series includes both base and instruction-tuned models, designed for tasks such as text generation, coding, and answering complex questions. These models are available for both research and commercial use, offering high performance in natural language processing (NLP) tasks.

Stable Assistant
Stable Assistant is Stability AI’s chatbot powered by the Stable LM language model and Stable Diffusion. It assists users with content generation, including blog posts, scripts, and captions. Stable Assistant is integrated with various services for generating images, video, and audio, making it a versatile tool for creative professionals. The chatbot is also available as a bot on platforms like Telegram, where users can interact with it for quick, AI-generated responses.

Licensing and API Services
Stability AI offers flexible licensing options for its AI models, allowing users to self-host and integrate the models into their workflows. Non-commercial and community licenses are available for smaller users, while enterprise licenses offer full commercial access with dedicated support and custom billing. Stability AI also provides an API service, enabling developers to build applications using their models.