Unverified95% confidenceFactTime unknown
BERT的参数规模为3.4亿，GPT-3的参数规模为1750亿
1
Sources
95%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
GitHub 8000+ Star：最全LLM学习资源库深度解析
githubWangRongsheng
Related Entities
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22B74% similarVerifiedQwen3-235B总参数量为2350亿，采用MoE架构，每次前向推理实际激活的参数量约22B73% similarPartially Verified通义千问发布了Qwen 3.8 Max模型，参数规模高达2.4万亿69% similarUnverified据Artificial Analysis数据，Qwen 3.8的270亿参数模型在Agentic指数上拿到51分，超过GPT-5.6 Terra、DeepSeek v4 Pro、GPT-5.6 Luna以及Gemini 3.7 Flash66% similarUnverified2.4万亿参数模型在BF16精度下考虑KV Cache和激活值后实际约需1.4-1.5TB显存66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/34539API
curl https://kongchang.com/api/v1/knowledge/claims/34539MCP
get_claim(id=34539)