Expired85% confidenceFactTime unknown
智谱发布了GLM 5.1 High Speed API，输出速度达到每秒400 token
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
GLM 5.1满血旗舰模型实测400 TPS，两分钟从草图到完整应用
bilibiliAI冷科长
Related Entities
Related Claims
Unverified在4×RTX 3090 SXM4配置下，GLM-5.2量化版本的Token生成速度约为6.3-6.8 tokens/秒，提示评估速度约为10.6 tokens/秒70% similarUnverifiedGrok 4.5处于快速模型档位，能达到约80 TPS（每秒Token数）的输出速度68% 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）66% similarUnverifiedGroq的LPU主打超低延迟推理，已实现超过500 Token/秒的输出速度66% similarUnverified本地部署DS4模型在STM32嵌入式开发测试中平均输出速度约为23 tokens/秒，推理时内存占用飙升至110GB，峰值速度达83.8 tokens/秒64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/6843API
curl https://kongchang.com/api/v1/knowledge/claims/6843MCP
get_claim(id=6843)