400 tokens/s: StepFun Step 3.7 Flash cuts Agent task costs

400 tokens/秒！阶跃Step 3.7 Flash，把Agent任务成本打到Claude零头

StepFun released Step 3.7 Flash, a sparse MoE model with 196B parameters plus a 1.8B ViT, activating 11B parameters per inference and reaching up to 400 tokens per second.

Why it matters: HKR-H/K/R all pass with concrete speed and parameter numbers. The feed does not disclose pricing, benchmark setup, or open-source terms, so this stays in the 78–84 quality update band.