Reflection AI unveils Beam, its first open-weight model
A 501B-parameter Mixture-of-Experts with 23B active, aimed at coding, reasoning and agents.
Reflection AI, the New York lab founded by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, announced Beam on October 5, 2026. It is the company's first frontier model and the first in a planned series.
Beam is a sparse Mixture-of-Experts model with 501 billion parameters, of which about 23 billion are active for each token. It was pretrained on 23.8 trillion tokens and then put through what Reflection calls one of the largest reinforcement learning runs by any open lab.
Reflection says Beam is competitive with GLM 5.2 and approaching Qwen 3.8 Max on coding and agentic tasks, while Kimi K3 remains ahead on raw capability. Its main claim is efficiency: comparable reasoning scores to GLM 5.2 with three to four times less inference compute.
The model is in final red-teaming. Early access is through a waitlist, and Reflection promises the weights, a technical report and a model card under the Apache 2.0 license later in October.
Takeaway
Beam is a serious US entry in the open-weight race, but for now only waitlisted users can try it.