Beam customization recipes are live for fine-tuning and reinforcement learning

Reflection AI’s developer docs describe recipes for adapting Beam with LoRA, full fine-tuning or reinforcement learning, or using it as a teacher to train smaller models. The docs say each recipe records its scripts, configurations, hardware and runtime. Reflection AI says the recipes are live on its website and in its cookbook repo, alongside Beam support across reinforcement learning and supervised fine-tuning stacks.