2 tracked versions of the step-3 family. Timeline ordered newest → oldest. Flagship version marked.
Step 3.5 Flash is StepFun’s open-weight text reasoning model for coding, tool use and high-throughput workloads. It remains available alongside the multimodal Step 3.7 Flash and the newer Step 5 Preview.
Step 3.7 Flash is an open-weight vision-language MoE model with 198B total parameters and about 11B active per token. It combines a 196B language backbone with a 1.8B vision encoder, supports a 256K context and offers low, medium and high reasoning settings.
See every provider hosting StepFun (阶跃星辰) models at /provider/stepfun.