Organisations form the backbone of the AI ecosystem, and understanding who builds what helps you contextualise each model’s purpose, capabilities, and reliability.
What an organisation page includes
Each organisation page provides an overview of the company or research group, including:| Section | Description |
|---|---|
| Overview | Basic details such as name, description, country, founding year, and website. |
| Published Models | All models released by the organisation, including archived or deprecated ones. |
Example organisation page
Types of organisations on AI Stats
| Type | Example | Description |
|---|---|---|
| Research labs | OpenAI, Anthropic, DeepMind | Focused on cutting-edge AI model development and innovation. |
| Tech companies | Google, Meta, Microsoft | Deploy large-scale AI infrastructure and consumer applications. |
| Startups | Mistral, Perplexity, Suno | Smaller teams building specialised or efficient AI systems. |
| Open-source communities | Hugging Face, Stability AI | Decentralised collaborations offering transparent model development. |
Provider vs organisation
It’s important to distinguish between organisations and API providers:| Aspect | Organisation | API Provider |
|---|---|---|
| Definition | The creator or maintainer of the model. | The platform offering API access to that model. |
| Example | Anthropic built the Claude series. | Anthropic, AWS and Google Vertex all provide access to Claude models. |
| Focus | Research, training, and model architecture. | Hosting, routing, pricing, and user experience. |
Contributing organisation data
If you notice missing or outdated information about an organisation, you can update it directly through GitHub.Contributions help maintain accuracy across the platform.
Contribute Organisation Data
Learn how to update or add organisation details safely.
Next steps
Now that you understand how organisations are represented, you can dive deeper into their benchmarks and comparative performance.Explore Benchmarks
See how different organisations’ models perform across standardised tasks.