The Bottom Line: A generative AI company known for the Stable Diffusion family of open-weight image models, now repositioned as a provider of purpose-built AI tools for professional creatives in music, gaming, and entertainment. Its Core Models span image (Stable Diffusion 3.5 Medium/Large/Turbo, SDXL Turbo), video (Stable Video Diffusion), 3D (Stable Fast 3D, SPAR3D), audio (Stable Audio 3.0 Small/Medium, with Large via API/enterprise), and language (Stable LM 2). Under the Stability AI Community License, models are free for research and commercial use for organizations with annual revenue below $1M; above that threshold an Enterprise license is required. Stable Audio 3.0 models are trained on fully licensed data, and users own their outputs. The company underwent a leadership change in 2024, restructured its balance sheet, and closed a $76M Series B in August 2026 backed by Sony Music, Universal Music, Warner Music, and Electronic Arts. SOC 2 Type II and SOC 3 compliance achieved.

Pricing model: Freemium — open-weight models are free to self-host and use commercially below $1M annual revenue under the Community License. Above that threshold, an Enterprise license applies. API access is credit-based: 1 credit = $0.01, with Stable Diffusion 3.5 Large at 6.5 credits per generation, Stable Audio 3.0 at 8 credits, and Stable Fast 3D at 10 credits . A separate API subscription tier starts at $10/month for 6,500 credits, though this is not the only access route .

Platforms: Web (API platform and developer docs). No native desktop or mobile app for the core generation models is documented.

Pros & Cons

Pros: Open-weight releases (Stable Diffusion 3.5, Stable Audio 3.0 Small/Medium) allow self-hosting, fine-tuning with LoRA, and full data control, which is rare among frontier model providers. The Community License is genuinely permissive—free commercial use below $1M annual revenue, with a clear upgrade path to Enterprise. The product line spans image, video, 3D, audio, and language under one roof, giving creative teams a single vendor for multiple modalities. Stable Audio 3.0 Small, Small-SFX, and Medium are downloadable from Hugging Face under the Community License, with user ownership of outputs . Enterprise deployment options include Azure AI Foundry, Amazon Bedrock, and NVIDIA NIM microservices. SOC 2 Type II and SOC 3 compliance suit enterprise procurement.

Cons: The company's identity and strategy have shifted significantly—from open-source-first image lab to a vertically focused creative tools provider backed by music labels and game publishers, which may alienate segments of its original open-source community. Stable Audio 3.0 Large, the most capable audio model, is not open-weight and requires API or an enterprise license, creating a tiered access structure . The $1M revenue threshold for the Community License means growing businesses face a licensing transition that requires legal and procurement attention. Historical instability (2024 debt and leadership crisis) and subsequent layoffs are part of the company's recent track record. The portfolio's breadth means individual models may not match specialist competitors in depth for any single modality.

Who is it for?

Ideal for: Developers and researchers who want open-weight generative models they can self-host, fine-tune, and integrate without per-generation API costs (below $1M revenue); professional creatives in music, gaming, and entertainment exploring AI-assisted workflows with licensed training data (Stable Audio 3.0); enterprises requiring SOC 2 compliance and cloud deployment options (Azure, AWS, NVIDIA NIM); teams that need multiple modalities (image, audio, 3D) from a single vendor rather than stitching together separate providers.