Stability AI

Stability AI

Overview

Stability AI is a London-based generative AI company best known for the Stable Diffusion family of open-weight image generation models — among the most influential open-source AI releases of the past five years. Founded in 2019 and brought to mass attention by the August 2022 release of Stable Diffusion 1.x, the company effectively launched the open-weight image generation ecosystem and seeded an entire industry of fine-tunes, custom UIs (Automatic1111, ComfyUI), and downstream applications. Stable Diffusion 3.5 (October 2024) is the most recent major SD release; the company has not yet shipped Stable Diffusion 4 as of May 2026.

Through 2025–2026, Stability AI has expanded beyond image generation into video (Stable Video Diffusion → SV4D 2.0), 3D asset generation, real-time enterprise rendering, and adjacent generative-media verticals. The company has weathered significant turbulence — a leadership transition with founder Emad Mostaque's 2024 departure, restructuring rounds, and intense competition from FLUX (Black Forest Labs), Midjourney, Recraft, and the major-lab image models — but remains the central node of the open-weight image generation ecosystem and a foundational reference point for downstream model fine-tuning.

Key Details

  • Founded: 2019
  • Founder: Emad Mostaque (departed 2024)
  • Headquarters: London, UK
  • Distribution Model: Stability AI Developer Platform + open-weight downloads (Hugging Face) + commercial license tiers
  • Strategic Positioning: Open-weight image / video / 3D generation; foundational reference for the broader open-weight generative-media ecosystem
  • Website: https://stability.ai

Current Models

  • Stable Diffusion 3.5 (Large / Large Turbo / Medium) — Most recent major SD release (October 2024); Large is 8B parameters, Medium is 2.5B (the deployment-preferred variant for many commercial integrations)
  • Stable Video Diffusion / SV4D 2.0 — Updated video generation model with improved real-world video performance
  • Stable Audio — Audio generation in the company's expanded portfolio
  • Stable Point Aware 3D / 3D Asset Models — Beyond image into 3D generation

Recent Developments

  • No Stable Diffusion 4 Yet (as of May 2026): Despite competitive pressure from FLUX 1.2 Pro Ultra, Midjourney V8.1, and frontier-lab image models (Nano Banana 2, GPT-5 image), Stability AI has not released SD 4 — the SD 3.5 family remains current.
  • Beyond Image Generation: Strategic expansion into video (SV4D 2.0), 3D asset creation, and real-time enterprise rendering — broadening the company's surface area beyond the saturated image generation tier.
  • SD 3.5 Medium Adoption: The 2.5B-parameter Medium variant has become the deployment-preferred option for many commercial integrations through 2025–2026, balancing quality and cost.
  • Ecosystem Centrality: Despite competitive pressure on flagship benchmarks, Stable Diffusion remains the foundational model referenced by the vast majority of open-weight image fine-tunes, control networks, and downstream applications — preserving Stability's gravitational role in the open-weight image ecosystem.

Why They Matter

Stability AI's historical importance as the lab that opened the floodgates on open-weight image generation is undimmed even as the company's competitive position has eroded. Stable Diffusion remains the foundational reference model for the entire open-weight image-generation ecosystem — every downstream fine-tune, ControlNet, LoRA, and custom UI assumes SD as the substrate. That ecosystem position is hard to dislodge even as FLUX and Midjourney push raw quality further. Strategically, Stability's 2025–2026 expansion into video, 3D, and enterprise-rendering verticals is a bet that the company can build durable revenue around its ecosystem position rather than continuing to compete on flagship image-quality benchmarks alone. Whether that bet pays off — and whether SD 4 ever arrives — will determine whether Stability AI remains a central player or becomes a foundational legacy referenced primarily by the labs and creators that built on top of it.

Last Updated

May 7, 2026

→ Back to Companies