According to the most recent Output Speed Ranking unveiled by Artificial Analysis, a prominent large model evaluation platform, Step 3.7 Flash—an open-source foundational model developed by StepFun (Jieyue Xingchen)—has clinched the top spot among mainstream models, boasting an impressive output speed of 409 tokens per second. Additionally, this model outperforms its counterparts in crucial metrics, including end-to-end response time, intelligence efficiency, and cost-effectiveness.
