Expired50% confidenceFactExact time
The optimized Qwen 397B model introduced by Meta achieved a level close to Claude Ops 4.8 through a self-trace approach.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/8/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/6/2026(expired)
Sources
GPT-5.6发布：Soul/Terra/Luna三层布局，编排能力训入模型
bilibili杨博士说AI6/29/2026
Related Claims
VerifiedGemma 4's 26B MoE model uses a Mixture of Experts architecture, with 26 billion total parameters but only approximately 4 billion parameters active at any given moment.58% similarUnverifiedQwen3's 235B-parameter MoE flagship model approaches or surpasses closed-source top-tier models on multiple authoritative benchmarks.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/248702API
curl https://kongchang.com/api/v1/knowledge/claims/248702MCP
get_claim(id=248702)