Generative image
3 labs · 30 models and products Theme: Companies · Generated from this page's data. Click a name to open its panel on the Labs page. Records resting on one source are tagged "one source".Generative image AI creates pictures from a written description, called a prompt. Stability AI's SDXL, for example, is a "diffusion" model, which learns to turn random noise into an image step by step. Stability AI builds open-weight models for images, video and audio, including Stable Diffusion, for media, games and music companies (per the site); its papers on SDXL and Stable Diffusion 3 describe how they work. Recraft makes tools that generate brand-consistent graphics, vector images and illustrations for designers (per the site); it says its Recraft V3 model took first place on a public text-to-image leaderboard in 2024.
Context
- Stability AI described SDXL, a model that generates images from text. 4 July 2023 arxiv.org/abs/2307.01952
We present SDXL, a latent diffusion model for text-to-image synthesis.
- Stability AI's Stable Diffusion 3 paper set out a "rectified flow" method linking noise and images in a straight line. 5 March 2024 arxiv.org/abs/2403.03206
Rectified flow is a recent generative model formulation that connects data and noise in a straight line.
- Recraft reports its V3 model took first place on the Artificial Analysis text-to-image leaderboard on Hugging Face. 30 October 2024 recraft.ai/blog/recraft-introduces-a-revolutiona
It secured #1 place with ELO rating of 1172.
- The government consulted on how UK copyright law should apply to training AI models. 17 December 2024 gov.uk/government/consultations/copyright-and-ar
This consultation seeks views on how the government can ensure the UK’s legal framework for AI and copyright
- The government published a report and impact assessment on copyright and AI under the Data (Use and Access) Act. 18 March 2026 gov.uk/government/publications/report-and-impact
A report and impact assessment on the use of copyright works in the development of artificial intelligence (AI) systems.
Key papers
- SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis arXiv preprint · 4 July 2023
- Scaling Rectified Flow Transformers for High-Resolution Image Synthesis (Stable Diffusion 3) arXiv preprint · 5 March 2024
Labs Round 8 (31 Oct 2026): introduction and facts as the researchers wrote them; "(per the site)" marks a sentence resting on this site's own description of that lab.
| Record | What it is | Place | Links | Confidence |
|---|---|---|---|---|
| Brand StudioChips, models and products | Stability AI's creative production platform that generates content in a company's own brand style. released (product)Key facts Announced 8 Apr 2026 (news index).Made by Stability AI · 2026-04-08 | Los Angeles (via Stability AI) | stability.ai/news-updates | one source |
| DeepFloyd IFChips, models and products | Text-to-image model from Stability AI's DeepFloyd research lab, built in three stages and able to write legible text inside images. released (research licence)Key facts Zero-shot FID 6.66 on COCO; T5-XXL-1.1 text encoder; trained on a custom LAION-A set of 1B image-text pairs; stage 1 at 64x64 (400M, 900M, 4.3B parameters), stage 2 to 256x256 (400M, 1.2B), stage 3 to 1024x1024 (700M, not yet released at launch); access: research licenceMade by Stability AI · 28 Apr 2023 (year from page image timestamps) | London | stability.ai/news-updates/deepfloyd-if-text-to-i | one source |
| Exactly.aiCompanies · also on Start-ups | Lets illustrators and artists train generative image models on their own work.Why it is here: reason: London platform on which artists train their own image-generation models on consented work | London | uktech.news/ai/ai-art-exactly-ai-funding-2024053 | one source |
| Gemini 2.5 Flash ImageChips, models and products | A Gemini image generation and editing model; the models page brands Gemini Image as "Nano Banana". releasedKey facts Theme: Generative image. Image-editing upgrade in Gemini August 2025 (index); model card "Gemini 2.5 Flash and Gemini 2.5 Flash Image" updated 26 Sep 2025. Successors: Nano Banana Pro (Nov 2025), Nano Banana 2 (Feb 2026).Made by Google DeepMind · 2025-08 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/models/model-cards/ · deepmind.google/blog/page/6/ | lab's own pages |
| Haiper Image 1.5 and Haiper Image 2.xChips, models and products | Haiper's text-to-image models, offered alongside its video models. released (API)Key facts Image 2.x priced $0.04 (720p), $0.08 (1080p), $0.12 (4K) per image; Image 1.5 priced $0.02 (720p) and $0.04 (1080p); text-to-image endpoint resolutions 720, 1080 and 2160; access: APIMade by Haiper · — (release dates not stated) | London | docs.haiper.ai/api-reference/endpoint/2-0-text-t · docs.haiper.ai/api-reference/api-pricing | one source |
| Imagen 2Chips, models and products | Second version of Google's Imagen text-to-image model, with an added model for judging image aesthetics. superseded (released)Key facts Released 2023; improved outputs via a specialised image aesthetics model; successors Imagen 3 and 4 already listed (BIZ-LAB-7164, 7165)Made by Google DeepMind · 2023 (exact day not stated on source) | London | deepmind.google/blog/2023-a-year-of-groundbreaki | one source |
| Imagen 3Chips, models and products | Google DeepMind's Imagen 3 image generation model, announced with Veo 2. released (superseded)Key facts Theme: Generative image. Announced with Veo 2 in December 2024 (index). SynthID image watermarking was first offered to Vertex AI customers using Imagen (29 Aug 2023). Successor: Imagen 4 (card 20 May 2025).Made by Google DeepMind · 2024-12 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/blog/page/7/ | one source |
| Imagen 4Chips, models and products | Google DeepMind's Imagen 4 image generation model. releasedKey facts Theme: Generative image. Model card dated 20 May 2025 (same month as I/O post "Fuel your creativity with new generative media models and tools"). Imagen is not listed on the models page as of 30 Oct 2026, where Gemini Image (Nano Banana) leads image generation.Made by Google DeepMind · 2025-05-20 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/models/model-cards/ | one source |
| Nano Banana 2 (Gemini 3.1 Flash Image)Chips, models and products | A fast Gemini image generation model with Pro-level features. releasedKey facts Theme: Generative image. Released 26 Feb 2026; model card "Gemini 3.1 Flash Image" same date. Follow-ons: Nano Banana 2 Lite / Gemini 3.1 Flash-Lite Image (30 Jun 2026); Nano Banana 2.1 (card 6 Oct 2026).Made by Google DeepMind · 2026-02-26 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/blog/rss.xml · deepmind.google/models/model-cards/ | lab's own pages |
| Nano Banana 2 Lite (Gemini 3.1 Flash-Lite Image) and Nano Banana 2.1Chips, models and products | Lighter and updated versions of the Nano Banana 2 image model. releasedKey facts Theme: Generative image. Nano Banana 2 Lite released 30 Jun 2026 with Gemini Omni Flash; card "Gemini 3.1 Flash-Lite Image" updated 30 Jun 2026; Nano Banana 2.1 card updated 6 Oct 2026.Made by Google DeepMind · 2026-06-30 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/blog/rss.xml · deepmind.google/models/model-cards/ | lab's own pages |
| Nano Banana Pro (Gemini 3 Pro Image)Chips, models and products | Google DeepMind's Gemini 3 Pro image generation and editing model. released (Gemini app and API)Key facts Theme: Generative image. Announced November 2025 ("Introducing Nano Banana Pro"; "Build with Nano Banana Pro, our Gemini 3 Pro Image model"); model card updated 20 Nov 2025. AI image verification added to the Gemini app the same month. Earlier image model: Gemini 2.5 Flash Image (card 26 Sep 2025).Made by Google DeepMind · 2025-11-20 | London (EC2M 4AG) (via Google DeepMind) | deepmind.google/blog/page/5/ · deepmind.google/models/model-cards/ | lab's own pages |
| RecraftCompanies · also on Start-ups | AI tools that generate brand-consistent graphics, vector images and illustrations for designers.Profile
Labs Round 8 deep profile (31 Oct 2026), as the researchers wrote it; confidence: one source. Sources Models and products (5)Version chainsRecraft V2 → V3 → V4 → V4.1 → V4.1 Flash
Key facts and news (8)
News
Labs Round 8 (31 Oct 2026), from the researchers' rows; one-source items are marked. | London (WC2B 5AH ) | Companies House · uktech.news/ai/recraft-ai-graphic-design-2024011 · find-and-update.company-information.service.gov. · recraft.ai/blog/recraft-introduces-a-revolutiona · recraft.ai/news/series-b-announcement | verified |
| Recraft V2 (Recraft 20B)Chips, models and products | Recraft's image model pitched as the first built for designers, with about 20 billion parameters. superseded (product)Key facts Announced 13 Mar 2024; about 20 billion parameters (post compares SDXL and Playground v2.5 at 2.6 billion each); user study on 1,632 PartiPrompts with 1,435 assessors, stated comparable to DALL-E 3 and ahead of Midjourney v6; access: productMade by Recraft · 2024-03-13Versions Recraft V2 → V3 → V4 → V4.1 → V4.1 Flash | London | recraft.ai/blog/recraft-v2 | one source |
| Recraft V3Chips, models and products | Recraft's 2024 image model for designers, able to place text in images and produce vector art. earlier model (still listed in docs; V4 is now the default)Key facts Ranked #1 on Artificial Analysis's text-to-image leaderboard with an ELO of 1172 (30 Oct 2024, company post)Made by Recraft · 2024-10-30 | London (WC2B 5AH ) (via Recraft) | recraft.ai/blog/recraft-introduces-a-revolutiona · recraft.ai/blog-announcements | lab's own pages |
| Recraft V4 (V4, V4 Pro, Vector, Vector Pro)Chips, models and products | Recraft's rebuilt image model family, including versions that output editable vector (SVG) files from a text prompt. liveKey facts V4 1024×1024 px about 10 seconds per image; V4 Pro 2048×2048 px about 28 seconds; vector models output layered SVG (17 Feb 2026)Made by Recraft · 2026-02-17 | London (WC2B 5AH ) (via Recraft) | recraft.ai/blog/introducing-recraft-v4-design-ta · recraft.ai/blog-announcements | lab's own pages |
| Recraft V4.1 (incl. Utility, Vector, Pro and Flash)Chips, models and products | Recraft's latest family of nine image models, from fast drafts to high-resolution and editable vector output. liveKey facts Released 14 May 2026; V4.1 $0.035 per image, Pro $0.21, Flash $0.007 (API pricing); V4.1 Flash median 1.3 seconds prompt-to-image (23 Sep 2026); V4.1 Utility Pro fifth on the Artificial Analysis Image Arena and highest outside Google and OpenAI (30 May 2026)Made by Recraft · 2026-05-14 | London (WC2B 5AH ) (via Recraft) | recraft.ai/docs/recraft-models/recraft-v4-1 · recraft.ai/press-releases/recraft-v4-1-utility-p | lab's own pages |
| Recraft V4.1 FlashChips, models and products | Recraft's fastest image model, producing a picture about 1.3 seconds after the prompt. released (product and API)Key facts Announced 23 Sep 2026; about 1.3 seconds median from prompt to image; in Recraft's 100-prompt test (21 Sep 2026) median 1.4 s and p95 1.8 s; "Enhance" regenerates with V4.1 Pro at 2048x2048; access: product and APIMade by Recraft · 2026-09-23 | London | recraft.ai/blog/meet-recraft-v4-1-flash | one source |
| SDXL (Stable Diffusion XL)Chips, models and products | Larger Stable Diffusion model for higher-resolution images, with a second text encoder and an optional refining model. superseded (open weights)Key facts Paper arXiv 2307.01952 submitted 4 Jul 2023; UNet backbone three times larger than earlier versions; second text encoder; separate refinement model; SDXL 1.0 announcement page not opened; access: open weightsMade by Stability AI · 2023-07-04 (paper)Versions Stable Diffusion v1 → 2.0 → SDXL → SDXL Turbo → SD3 → SD3.5 | London | arxiv.org/abs/2307.01952 | one source |
| SDXL Turbo (Adversarial Diffusion Distillation)Chips, models and products | Fast version of SDXL that makes an image in a single step, close to real time. released (non-commercial open weights)Key facts Steps cut from 50 to one; 512x512 image in 207 ms on an A100 (67 ms for one UNet pass); method paper arXiv 2311.17042 submitted 28 Nov 2023; access: weights and code on Hugging Face under non-commercial research licence, demo on Clipdrop; listed among Stability Core Models (page updated 20 May 2026)Made by Stability AI · 2023-11-28Versions Stable Diffusion v1 → 2.0 → SDXL → SDXL Turbo → SD3 → SD3.5 | London | stability.ai/news-updates/stability-ai-sdxl-turb · arxiv.org/abs/2311.17042 | lab's own pages |
| Stability AICompanies | Builds open-weight generative models that make images, video and audio from text, including Stable Diffusion and Stable Audio, for media, games and music companies.Profile
Labs Round 7 deep profile (30 Oct 2026), as the researchers wrote it; confidence: single_source. Sources Models and products (27)Version chainsStable Audio 2.0 (3 Apr 2024) → Stable Audio Open (5 Jun 2024) → Stable Audio Open Small (14 May 2025) → Stable Audio 2.5 (10 Sep 2025) → Stable Audio 3.0 (20 May 2026)StableLM → Stable LM 2Stable Diffusion v1 → 2.0 → SDXL → SDXL Turbo → SD3 → SD3.5
Selected papers (12)
Key facts and news (7)
News
Labs Round 8 (31 Oct 2026), from the researchers' rows; one-source items are marked. | Los Angeles | Companies House · stability.ai/news-updates/stability-ai-latest-fu · tech.eu/2026/08/25/stabiityai-raises-76m-with-wa · stability.ai/news/stability-ai-secures-significa · find-and-update.company-information.service.gov. | verified |
| Stability AI SolutionsChips, models and products | Stability AI's packaged generative AI services for enterprise creative production. released (enterprise service)Key facts Announced 5 Aug 2025 (news index). Related: SOC 2 Type II and SOC 3 compliance announced 4 Aug 2025.Made by Stability AI · 2025-08-05 | Los Angeles (via Stability AI) | stability.ai/news-updates | one source |
| Stable AssistantChips, models and products | Stability AI's assistant product, through which its image models were offered to users. released (product)Key facts Feature release announced 9 Jul 2024; SD3 Medium offered on free three-day trials of Stable Assistant (12 Jun 2024).Made by Stability AI · 2024-07-09 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |
| Stable CascadeChips, models and products | A Stability AI model announced in February 2024; launch post not opened, so details are not recorded. released (model)Key facts Announced 12 Feb 2024 (product news index).Made by Stability AI · 2024-02-12 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |
| Stable Code 3BChips, models and products | A small (3-billion-parameter, per its name) coding language model from Stability AI; launch post titled "Coding on the Edge". released (model)Key facts Theme: coding model (no matching theme tile). Announced 16 Jan 2024. Follow-up: Stable Code Instruct 3B (25 Mar 2024).Made by Stability AI · 2024-01-16 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product · stability.ai/news-updates/category/Product | one source |
| Stable Diffusion (first public release, v1)Chips, models and products | The original open text-to-image model that made pictures from typed prompts on an ordinary consumer graphics card. superseded (open weights)Key facts Runs on under 10 GB of VRAM; 512x512 images in a few seconds; trained on LAION-Aesthetics, a subset of LAION 5B; over 10,000 beta testers creating 1.7 million images a day; the post credits CompVis latent diffusion work, LAION and EleutherAI, and Runway's Patrick Esser as project lead; access: open weights (code on GitHub, model card on Hugging Face)Made by Stability AI | London | stability.ai/news-updates/stable-diffusion-annou | one source |
| Stable Diffusion 2.0Chips, models and products | Second major Stable Diffusion release, with a new text encoder, higher default resolution and a 4x upscaler. superseded (open weights)Key facts Default resolutions 512x512 and 768x768; upscaler raises resolution 4x; OpenCLIP text encoder (LAION, supported by Stability AI); trained on an aesthetic subset of LAION-5B filtered for adult content; access: open source plus Stability API and DreamStudioMade by Stability AI | London | stability.ai/news-updates/stable-diffusion-v2-re | one source |
| Stable Diffusion 3 (incl. SD3 Medium)Chips, models and products | Stability AI's third-generation text-to-image model; the 2-billion-parameter Medium version was released with open weights. released (API; open weights for Medium)Key facts Announced 22 Feb 2024; API available 17 Apr 2024; SD3 Medium open weights 12 Jun 2024 (2 billion parameters; Hugging Face under a community licence, separate licence for large-scale commercial use; page edited 5 Jul 2024). Licence update post 5 Jul 2024. Succeeded by Stable Diffusion 3.5 (22 Oct 2024).Made by Stability AI · 2024-02-22Read the paper | Los Angeles (via Stability AI) | stability.ai/news-updates/stable-diffusion-3-med · stability.ai/news-updates/category/Product | lab's own pages |
| Stable Diffusion 3.5Chips, models and products | Stability AI's text-to-image model family released in October 2024 and later offered through cloud marketplaces. released (model)Key facts Announced 22 Oct 2024; ControlNets for SD 3.5 Large 26 Nov 2024; SD 3.5 Large on Amazon SageMaker JumpStart 14 Nov 2024, Amazon Bedrock 19 Dec 2024, Microsoft Azure AI Foundry 12 Feb 2025; TensorRT-optimised versions "2X Faster Performance and 40% Less Memory" on NVIDIA RTX (12 Jun 2025); SD 3.5 NIM with NVIDIA (12 Aug 2025). Launch page could not be opened.Made by Stability AI · 2024-10-22 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product · stability.ai/news-updates | one source |
| Stable Fast 3DChips, models and products | A model that makes a 3D asset from a single image quickly. released (model)Key facts Announced 1 Aug 2024. Followed by Stable Point Aware 3D (8 Jan 2025).Made by Stability AI · 2024-08-01 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |
| Stable Point Aware 3D (SPAR3D)Chips, models and products | A model that generates a complete 3D object from an image and allows real-time editing. released (model)Key facts Announced 8 Jan 2025 (product news index).Made by Stability AI · 2025-01-08 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |
| Stable Video 3DChips, models and products | A model that generates new viewing angles and 3D objects from one image. released (model)Key facts Announced 18 Mar 2024 (product news index). Builds on the Stable Video Diffusion line.Made by Stability AI · 2024-03-18 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |
| Stable Zero123 (and Stable Zero123C)Chips, models and products | Model that generates new views of an object from a single picture, as a step to making 3D objects. released (non-commercial; Zero123C commercial for members)Key facts Training about 40 times faster than Zero123-XL; based on Stable Diffusion 1.5; trained on a filtered subset of Objaverse; 24 GB VRAM recommended for 3D generation; Stable Zero123C (CC-BY and CC0 objects only) usable commercially with a Stability membershipMade by Stability AI · 13 Dec 2023 (post updated 8 Jan 2024) | London | stability.ai/news-updates/stable-zero123-3d-gene | one source |
| TripoSRChips, models and products | A model that makes a 3D object from a single image quickly. released (model)Key facts Announced 5 Mar 2024 (product news index).Made by Stability AI · 2024-03-05 | Los Angeles (via Stability AI) | stability.ai/news-updates/category/Product | one source |