Reflection AI

Reflection AI builds open large language models using reinforcement learning for autonomous coding and agentic reasoning tasks.

Reflection AI develops open foundation models trained with reinforcement learning techniques. These models handle complex software engineering by planning code, debugging errors, and executing full tasks without human input. The company releases model weights publicly to support research and applications in AI agents.