# Fireworks AI

> A firm that runs open AI models for you, so you need no chips.

**Field:** Artificial Intelligence  
**Category:** Models & Providers  
**Also known as:** Fireworks, fireworks.ai

## Think of it like

It is like a kitchen you can rent. Bring any open recipe, and its cooks make it fast.

## Definition

Fireworks AI is an inference provider that runs open models for you through one API. You can start with no setup and pay per token. Later you can rent your own set of chips as your use grows. It serves well-known open models from many labs, and it can run your own. It also offers fine-tuning, and its servers run in more than 30 regions.

## Example

A team wants an open model for its app but owns no GPUs. It sends its prompts to a GLM model on Fireworks AI and pays per token.

## Why it matters

Open models are only useful if you can run them. Providers like this one make that a single API call.

## Common mistake

Thinking it makes its own big models. It runs other labs' open models, and yours.

## Three principles

- **Runs open models.** Through one API.
- **Pay per token.** Or rent your own chips.
- **Fine-tuning too.** From LoRA to full training.

## Related terms

- [Inference Provider](https://learn.benibix.com/term/ai/inference-provider)
- [Together AI](https://learn.benibix.com/term/ai/together-ai)
- [Groq](https://learn.benibix.com/term/ai/groq)
- [Replicate](https://learn.benibix.com/term/ai/replicate)
- [Open-Weight Model](https://learn.benibix.com/term/ai/open-weight-model)
- [Low-Rank Adaptation (LoRA)](https://learn.benibix.com/term/ai/low-rank-adaptation)

## Taught in

- [AI Inference Providers](https://learn.benibix.com/learn/ai-models-providers/ai-inference-providers): AI Models and Providers, 12/13

![Two lanes split by a dashed line. Top lane, model hub: the hub holds model files and cards, and an arrow labelled download carries one to your computer, where you run it; Hugging Face is the example. Bottom lane, inference provider: your app sends a prompt to the provider's chips, an answer comes back, and you pay per token; Groq, Together AI and Replicate are the examples.](https://learn.benibix.com/diagrams/ai2-hub-or-provider.svg)

---

Source: https://learn.benibix.com/term/ai/fireworks-ai
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-19
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
