Featherless Llm

Featherless is a serverless hosting service for open large language models. With one API key you get access to more than 40,000 open models, from small efficient ones to very large flagship models. The goal is to let you test and deploy open models without managing GPUs yourself.

Model catalogue

The catalogue is organised into lists such as top language-specific models, top small models, top roleplay and creative writing, top reasoning, most popular, trending and top productivity. Examples shown on the site include DeepSeek, GLM, Kimi and Qwen models of different sizes. Each listing shows the model's size and usage, which helps you compare options. A tool called Simple Jev can turn any model into a classifier.

Plans

Featherless charges flat monthly prices rather than per token. All 40,000+ models are available on every plan, and the plans differ in context length, concurrency and intended use. The Chat plan, at 25 dollars a month, is for roleplay, stories and long conversations typed by you, with unlimited tokens, 32K context and four concurrent units, and it works with front ends like SillyTavern, RisuAI and Wyvern. A Developer plan is for coding agents and API traffic. Dedicated GPUs are also offered, and a cost calculator helps you estimate needs.

Who it suits

Featherless suits developers who want to try many open models through one API, hobbyists who use chat front ends with open models, and teams that prefer predictable monthly costs over usage billing.

Things to consider

Flat pricing with concurrency limits works well for steady use but may not suit heavy bursts, so check how many parallel requests your app needs. Model quality varies widely across a catalogue this large, so evaluate a few candidates on your own tasks. Each open model has its own licence, which you should check before commercial use. Review the data handling policy if you send sensitive prompts, and use the status page to judge reliability before relying on it in production.

The large catalogue also makes it a convenient place to compare models side by side before choosing one for a product.

Start on a single monthly plan and upgrade only when limits bite.