[Hugging Face (Twitter)]() RT [@HuggingPapers](): StepFun's Step 3.5 Flash A sparse MoE model with 196B parameters, 11B active per token. Achieves frontier-level reasoning comparable to GPT-5.2 xHigh and Gemini 3.0 Pro at 1/6th the decoding cost. Ranks on MathArena with 97.3% on AIME 2025.