Unverified50% confidenceBenchmarkExact time
Moonshot Kimi K2.5 has a total parameter count of 1.04T (32B active, 384 experts), and even with QAT-INT4 quantization it still requires 8×H200 to hold the weights.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/14/2026
First Seen
Valid until: 10/28/2026
Sources
Related Entities
Related Claims
VerifiedKimi K2.6 adopts a Mixture of Experts (MoE) architecture, with a total parameter count at the trillion level and only 32B parameters actually activated.74% similarUnverified月之暗面发布的Kimi K2参数量达到1T，率先跨入万亿参数门槛70% similarUnverifiedOn the B300 benchmark, Kimi K3's new optimization solution on vLLM achieved a 2.2 to 2.8× throughput improvement compared to version v0.27.1.70% similarUnverifiedKimi K3 is a 2.8T-parameter multimodal LatentMoE model that uses 896 experts with top-16 routing and performs routing computation in a 3584-dimensional latent space.69% similarUnverifiedOn BU Bench V1, the 4B-parameter SFT model surpassed the 45% score reported by DeepSeek V4 Pro and Kimi K2.6 under the open-source Browser Use framework.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1027764API
curl https://kongchang.com/api/v1/knowledge/claims/1027764MCP
get_claim(id=1027764)