Unverified50% confidenceBenchmarkExact time
Chinchilla模型拥有70B参数和1.4T tokens，在同等算力下显著优于拥有280B参数和300B tokens的Gopher模型
1
Sources
50%
Confidence
Long-term
Relevance
8/21/2026
First Seen
Sources
Related Claims
UnverifiedChinchilla模型以70B参数配合4倍数据量即可超越280B参数的Gopher模型性能78% similarUnverified在Gemma 4 12B模型的Token生成测试中，Spark达26.4 tokens/s、Halo 24.6 tokens/s、M4 Pro最快33.8 tokens/s70% similarUnverifiedChinchilla以700亿参数在多项基准上超越2800亿参数的Gopher69% similarUnverified万象模型提供了 1.3B 和 14B 参数量的不同规模版本64% similarUnverifiedMiniMax M3 每个 token 推理时约激活 23B 参数63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/781269API
curl https://kongchang.com/api/v1/knowledge/claims/781269MCP
get_claim(id=781269)