Unverified50% confidenceBenchmarkExact time
Ultrafast 能够以每秒最高 750 tokens 的速度生成内容
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/12/2026
First Seen
Valid until: 12/11/2026
Sources
Related Entities
Related Claims
Unverified该配置下35B模型实测单人最高速度约80 Tokens/秒,并发可达500 Tokens/秒以上75% similarUnverified在预填充速度测试中,Halo约650 tokens/s,Spark约2000 tokens/s,Spark快了将近三倍74% similarUnverifiedGrok 4.5每秒输出速度80 token,上下文窗口高达50万token(约等于40万个英文单词)74% similarUnverifiedGroq的LPU主打超低延迟推理,已实现超过500 Token/秒的输出速度70% similarUnverified70B参数模型(140GB权重)在1TB/s带宽下理论最快生成速度约7 tokens/秒69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/910293API
curl https://kongchang.com/api/v1/knowledge/claims/910293MCP
get_claim(id=910293)