StepFun launches Step 5 Preview, a 600B MoE model with 1M context for agentic work

StepFun has released Step 5 Preview, its new flagship sparse Mixture-of-Experts (MoE) model designed for software engineering, professional knowledge work, and finance. The model features 600 billion total parameters with 27 billion active per token, a 1M-token context window, and support for text, image, and video input.

• Architecture: 92 Transformer layers in a narrow-deep layout trained with on-policy long-horizon reinforcement learning.
• Benchmarks: Scores of 66.4 on FrontierFinance, 83.3 on DRACO, 67.7 on DeepSWE v1.1, and 44 on Artificial Analysis's Intelligence Index.
• Pricing: $1.00 per 1M input tokens and $2.70 per 1M output tokens, positioned as lower cost than comparable models.
• Availability: Hosted via API with open weights scheduled for October 15, 2026.

StepFun states the model delivers comparable intelligence at a substantially lower task cost, aiming to reach the Pareto frontier for agentic workloads.