Echo

The Distributed Reinforcement Learning Framework

Research paper: published

Models: open-sourced

As AI systems grow in scale and generality, RL has become a critical layer for alignment. However, most co-located RL frameworks face a deadlock: inference interrupts training, training delays inference.

Echo is Gradient’s distributed RL framework, running on everyday consumer devices.

Dual Swarm Architecture

Echo breaks the cycle by decoupling them into two independent swarms:

• Inference on distributed devices (via Parallax)

• Training on high-performance clusters

Echo Modes

We have open-sourced several early explorations with Echo, including:

• A 7B math model beating Qwen2.5–32B

• A 30B Sokoban model outperforming DeepSeek-R1 & GPT-OSS-120B

• A 32B LoRA logic model ≥99% accuracy

See all models in our Hugging Face Collection.

Learn More

Read the blog. Read the full research paper.