Verified65% confidenceFactExact time
Beam's total parameter count is 501B, with 23B parameters activated per token (about 4.6% of the network is activated).
3
Sources
65%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedBeam's VRAM requirement is close to that of a 501B-scale model, while its compute requirement is close to that of a 23B-scale model.77% similarUnverifiedReflection-Beam-1.0 has a total parameter scale of 501B, with only 23B parameters activated per inference.75% similarUnverifiedBeam adopts a Sparse Mixture-of-Experts (Sparse MoE) architecture with 52 layers, activating only 23 billion parameters at a time.65% similarUnverifiedMiniMax M3 每个 token 推理时约激活 23B 参数64% similarUnverifiedBeam's activation ratio is about 4.6%, lower than Mixtral 8x7B's roughly 46%.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011086API
curl https://kongchang.com/api/v1/knowledge/claims/1011086MCP
get_claim(id=1011086)