Expired50% confidenceFactTime unknown
K2.6 uses only approximately 3% of its total parameters per inference, making its inference cost comparable to that of a 32B dense model.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Kimi K2.6 Open-Source Hands-On: How Strong Is Its Orchestration of 300 Concurrent Agents?
bilibiliAi学习的老章
Related Claims
Unverified千问三235B MoE总参数量2350亿，但每次推理实际激活的参数约22B70% similarUnverifiedMoE模型中'35B'指总参数量，每个token只激活少量Expert（通常2-4个），实际计算量可能仅相当于6-7B的Dense模型68% similarUnverified千问3.6的35B MOE模型每次只激活其中3B的参数67% similarUnverifiedMoE models activate only a small fraction of total parameters per inference; Qwen3 Max may activate between 1/8 and 1/16 of the total parameters per inference.67% similarUnverifiedHy3 adopts a Mixture of Experts (MoE) architecture with a total parameter count of 295 billion (295B), activating only 21 billion (21B) parameters per inference.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/55585API
curl https://kongchang.com/api/v1/knowledge/claims/55585MCP
get_claim(id=55585)