Unverified50% confidenceBenchmarkExact time
Meta的Llama 3.1支持405B参数，与Qwen2.5、DeepSeek-V3等开源模型在多项基准测试中已接近甚至超越部分闭源商业模型
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/16/2026
First Seen
Valid until: 10/14/2026
Sources
Related Claims
UnverifiedQwen2.5-72B 在多项基准测试中接近 Llama-3.1-405B 的表现67% similarUnverifiedQwen3.5的参数激活率为4.3%，而DeepSeek V3的671B模型激活率为5.5%67% similarUnverified参与测试的四款模型是QDAM 3.7 Plus、LongCat 2.0、DeepSeek Feast Pro和MiniMax Mi366% similarUnverifiedQwen3-4B在多项推理类评测指标上已与DeepSeek V3或GPT-4o相差无几66% similarUnverified千问3的235B参数MoE旗舰版在多项权威基准测试中接近甚至超越闭源顶级模型65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/527889API
curl https://kongchang.com/api/v1/knowledge/claims/527889MCP
get_claim(id=527889)