Hugging Face is an open-source AI community and model-hosting platform founded in 2016, hosting hundreds of thousands of models and datasets on its Hub, with the Transformers library at 200,000+ GitHub stars, plus Inference API, Spaces and enterprise plans.
Google AI is Google's AI portfolio (founded 1998, headquartered in Mountain View, USA), centered on the natively multimodal Gemini models, the Vertex AI enterprise platform and in-house TPUs, with free tiers and usage-based APIs.
Anthropic is a leading AI Platform provider serving global customers.
Founded in 2022 and headquartered in San Francisco, Together AI is an inference cloud for open-source LLMs, offering low-latency APIs for 200+ models, LoRA fine-tuning and GPU clusters with an OpenAI-compatible API.
Fireworks AI is an ultra-fast LLM inference platform founded in 2022 in California, USA. Its FireAttention kernel pushes latency for open-source models like Llama and DeepSeek to millisecond levels, with an OpenAI-compatible API.
Replicate was founded in 2019 in San Francisco with Y Combinator backing, offering a cloud platform to run open-source AI models like image, video and audio via one API billed per second, with tens of millions of cumulative runs.