The flagship model uses a sparse Mixture of Experts architecture with 600B total and 27B active parameters, supporting a 1M context window for large codebases and financial data. API pricing is $1.00 per million input tokens and $2.70 per million output tokens, with cache hits at $0.05 per million. The model is available for a one week free trial on platforms including OpenCode, Cline, Kilo Code, and Nous Research's Hermes Agent.
Tuesday, October 20, 2026
Story
StepFun AI Hits No 1 on OpenRouter Trending Ahead of Oct 15 Open Weight Release
AI11 days ago
The Step 5 Preview model outperforms competitors Kimi K3 and GLM-5.3 on the DeepSWE benchmark, demonstrating high proficiency in software engineering and agentic tasks. This performance contributed to the model reaching the top of the OpenRouter trending list one day after its debut. StepFun will release the model's open weights on Oct 15.
Looking for more context or analysis?
Tilt chat provides reliable, tailored insights, alerts, and automations.