Overview

​

Free, fast & easy to use OSS LLM API

llmOS solves the following problems for devs:

• Slashing large OpenAI bills by 100% and making GPT-3.5 level LLMs completely free*.

• Takes care of deploying, scaling and optimizing OSS LLMs for production, which otherwise is hard and time-consuming.

• Upcoming: enables LLM devs & application builders to earn money and build on an open-source, permissionless LLM operating system.

*llmOS is still in alpha so every user has 500M tokens per month. This will increase as we are scaling up. Please write to us if you already need a higher limit.

​

LLM selection

The first model we serve is Mistral 7B Instruct model which is on bar with GPT-3.5 on many benchmarks.

We will be adding more LLMs. If you’d like to request a specific LLM, please write on Discord.

​

How can this be free? Open, permissionless peer-to-peer LLM network

Today, AI is like the closed intranets of the 90’s and owned by only a handful few large corporations. Moreover, there is regulatory capture to stifle open-source AI.

We believe AI should be more like the Internet: an open, permissionless and citizen owned network - that’s why we are building llmOS.

This is also how llmOS can be free - think of it as torrents for LLMs. We are in the early phases of building out the network.

Register your interest to directly participate in the network. We will reach out in the coming months as we open the network to early users.

​

Community