Reflection introduces US-originated open-weight model "Beam"
Reflection announced its first open-weight model "Beam" on the 5th (US time). It is a Mixture-of-Experts (MoE) model with 501B total parameters and 23B active parameters, designed for coding, reasoning, and agentic workloads.
US-based Reflection announced its first open-weight model, "Beam," on the 5th (US time). The model is a Mixture-of-Experts (MoE) with a total of 501 billion (501B) parameters and 23 billion (23B) active parameters, designed for coding, reasoning, and agentic workloads.
It is currently in the evaluation stage by red teams and has started early acceptance. Model weights, technical reports, and model cards are scheduled to be released within October.
This item is a summary. The full, checkable version is at the original source:
Read the original at Impress Watch ↗Read the original at MarkTechPost ↗