Foundational image models
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | territory of the AI industry |
|---|---|
| Categories | 13 |
| Companies | 65 |
| Part of | neural networks |
| Map last changed | 2026-08-29 |
| Map | mapofai.org/t/foundational image models |
Foundational image models is a territory of the AI industry on map of ai. In short: foundational image models generate images from text and edit existing images at scale. The map lists 65 products and companies in 13 categories.
Text-to-image generation
Main article: text-to-image generation
Text-to-image generation uses AI models to create images from natural language descriptions. Used by designers, marketers, content creators, and enterprises for rapid prototyping and asset generation.
Products and companies
- stable diffusion
- dall-e 3
- midjourney
- flux.1
- flux.2
- imagen 3
- gpt-image-1
- firefly
- ideogram 2
- playground v3
- qwen-image
- hugging face
- playground
Papers, last 30 days
- Breaking Visual Conventions: Spectrally Coupled Weight Perturbation for Creative Image Generation 2026-11-17
- UniWeaver: Harnessing Unified Multimodal Models for Interleaved Text-Image Generation 2026-11-10
- T2I Models Rarely Invert Negation but Often Fail to Suppress Negated Concepts: Benchmarking Negation Suppression in Text-to-Image Generation 2026-11-06
- Whose Values Are Generated? Measuring Visual Value Leakage in Multilingual Text-to-Image Generation 2026-11-03
- T-DIG: Text-Conditioned Diversity for Image Generation 2026-10-21
Open ground: Most models struggle with consistent character identity across multiple images and fine-grained control over specific visual elements without verbose prompting.
Open-source model hubs
Main article: ai open source model hubs
Platforms hosting open-source foundational image models for download and deployment. Used by researchers, developers, and organizations building computer vision applications without vendor lock-in.
Products and companies
Open ground: Most hubs lack integrated model evaluation frameworks that let users benchmark performance across their specific use cases before deployment.
Image editing & inpainting
Main article: image editing inpainting
Products and companies
Image api & inference platforms
Products and companies
Fine-tuning & personalization
Products and companies
- dreambooth
- lora training
- astria
- pictor
Portrait ai apps
Portrait AI apps use foundational image models to generate, edit, and enhance human faces and headshots. Used by content creators, professionals needing headshots, and social media users.
Products and companies
Open ground: Most portrait apps struggle with consistent identity preservation across multiple generations and custom styling without losing facial recognition accuracy.
Design & creative generation
Main article: design creative generation
Products and companies
Generative stock imagery
Generative stock imagery uses AI models to create photorealistic and artistic images on demand, replacing traditional stock photo libraries. Used by marketers, designers, content creators, and enterprises needing fast, scalable visual content.
Products and companies
Open ground: Most generators struggle with consistent brand identity across batches and lack seamless integration with existing design workflows and asset management systems.
Frontier image generators
Foundational image models are large-scale neural networks trained on broad image datasets that generate, understand, or manipulate images from text or other inputs. Product teams, researchers, and application developers use these models as base layers for downstream image generation tasks.
Products and companies
Open ground: OpenAI, Google, Anthropic, and Stability AI each ship models that generate images from text; none of them provides a model that generates video from text at comparable scale and latency for real-time interactive use.
Inference platforms
Inference platforms provide hosted compute infrastructure to run foundational image models at scale. Machine learning engineers, application developers, and enterprises use these platforms to generate, edit, or analyze images without managing their own GPU clusters.
Products and companies
Papers, last 30 days
- Flide: Accelerating Decision Inference with Exact State Forking 2026-10-30
- How Much Training-Inference Mismatch Can On-Policy Distillation Tolerate? 2026-11-08
- IDEA: Intent-Dominance Evidence Audit for Inference-Time Backdoor Detection in GUI Agents 2026-11-18
- Reducing Batch Makespan for LLM Agent Workloads through Inference-Environment Coordination 2026-11-12
- From Sequence Uncertainty to Phenotype Hypotheses: Uncertainty-Conditioned Inference over Biomedical Knowledge Graphs 2026-11-19
Open ground: Replicate, Together, Baseten, and Modal all offer on-demand inference APIs for image models; none of them yet offers guaranteed sub-100ms latency for real-time interactive image generation in browsers without client-side model loading.
Creative suites
Main article: creative suites
Foundational image models are large-scale neural networks trained on broad image datasets that generate, understand, or manipulate images; used by researchers, product teams, and developers building downstream applications.
Products and companies
Open ground: Stability AI, OpenAI, and Google each ship image generation models; none of them currently open-sources a foundational image model at the scale and capability of their deployed versions.
Image editing models
Image editing models are AI systems that modify, retouch, or transform existing images through techniques like inpainting, outpainting, and style transfer. Designers, photographers, content creators, and software developers use them to alter images without manual pixel-level work.
Products and companies
Papers, last 30 days
- TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing 2026-11-04
- EditAgent: Rethinking Agentic Image Editing 2026-10-29
- Text-Aware, Vision-Blind: How Multimodal Image Editing Systems Handle Unsafe Inputs 2026-11-19
- MLIE-CoT: Chain-of-Thought Reasoning via Multi-Level Image Editing 2026-11-18
Open ground: Stability AI, Adobe, and Runway each ship inpainting and outpainting via APIs or applications; none of them yet offers real-time collaborative multi-user editing on the same image canvas with live synchronization of AI-assisted changes.
3d and world models
Main article: 3d and world models
3D and world models generate or represent three-dimensional environments, objects, and scenes that can be viewed and navigated from multiple angles. Game developers, architects, VFX studios, and roboticists use these to create digital worlds, design spaces, and train autonomous systems.
Products and companies
Open ground: Nvidia, Stability AI, Meshy, and Spline each generate 3D models from text or images, but none of them generates fully articulated, physically simulated multi-object scenes with realistic material properties and lighting that persist across extended interactive navigation without manual refinement.
See also
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.