PANews reported on September 10th that Prime Intellect, a decentralized AI protocol, released its Reinforcement Learning Fine-Tuning (RFT) service, providing a full-stack RL infrastructure to empower AI-native companies to create and train custom agents. The service includes an RL-trainer, a validator, and an environment hub, enabling model training within a reinforcement learning environment. Additionally, Prime Sandboxes, an isolated Docker environment designed for secure code execution and LLM benchmarking, was introduced.

Earlier news reported that the decentralized AI protocol Prime Intellect completed US$15 million in financing, led by Founders Fund .