Skip to content
← Models
Open weightsStability AI/10 trustVerified
Stable Diffusion 3 logo

Stable Diffusion 3

MMDiT text-to-image model with rectified flows and strong typography.

Visit official page →

About

Stable Diffusion 3, released in June 2024, is a Multimodal Diffusion Transformer trained with rectified flow matching, using three fixed text encoders (OpenCLIP-ViT/G, CLIP-ViT/L, and T5-XXL). The 2-billion-parameter Medium release was designed to run on consumer hardware while improving photorealism, complex prompt understanding, and typography.

It shipped under the Stability Community License, which permits free research and commercial use for entities under $1 million in annual revenue, with a paid enterprise license above that threshold. The release was widely seen as Stability AI's attempt to reclaim ground against closed image models, and the later 3.5 line refined the same architecture.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

June 12, 2024

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

text-to-imagetypography-renderingcreative-designlocal-image-gen

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 82; Trending score 82/100 from researcher assessment.

More in the stable-diffusion family

Sources