Reflection claims Beam matches GLM-5.2 with 3-4x less inference compute

Reflection AI claims its Beam matches Z.ai's GLM-5.2 on advanced reasoning benchmarks while using 3-4x less inference compute. Beam is a text-only mixture-of-experts model with 501 billion total parameters and 23 billion active, versus roughly 744 billion total and 40 billion active for GLM-5.2.