From the source
Reflection AI released Beam, its first open-weight model, under an Apache 2.0 license on
Hugging Face.
Beam is a sparse Mixture-of-Experts model with 501 billion total parameters and 23 billion active, designed for coding, reasoning, and agentic workloads.
The release includes model weights in BF16, FP8, and NVFP4 precision, a speculative decoding drafter, and an ecosystem of over 40 partners and integrations.
