StepFun lists Step 5 Preview on OpenRouter on October 8, a 600B sparse MoE focused on enterprise agent workflows
Shanghai lab StepFun made Step 5 Preview public on OpenRouter on October 8, with 600B total and 27B active parameters, a 1M token context window, priced at $1.00 per million input and $2.70 output, with open weights planned for October 15.
Shanghai AI lab StepFun made Step 5 Preview public through OpenRouter on October 8. The model is built on a sparse Mixture-of-Experts architecture with 600B total and 27B active parameters, a 1M token context window, and a maximum output of 64K tokens. Pricing is $1.00 per million input tokens, $2.70 per million output tokens, and $0.05 per million for cache reads. StepFun is also counting down to an open-weights release on October 15. The API has been live since September on platforms such as Tencent Cloud TokenHub, and the OpenRouter listing extends distribution. OpenRouter-bundled benchmarks position Step 5 Preview as a competitive choice for software engineering, knowledge work, and financial analysis agent tasks. Open-source developer tools including Cline, OpenClaw, and Nous Research's Hermes Agent are already routing substantial token volumes to the model. StepFun's April-released Hy3 preview remains in pre-launch preview status. The OpenRouter listing reports around 38 tokens per second throughput and 3.8 second median latency as service-side reference data.