StepFun's Step 5 Preview beats Gemini 3.8 Flash on the AA Intelligence Index (44 vs 41) at ~42% lower cost per task. Free for a week on Nous Portal and OpenRouter, open weights due Oct 15.

Get the guide: link in bio.

#Step5

The Chinese lab's 600B MoE flagship scores 44 on the Intelligence Index at $1 input and $2.70 output, undercutting GPT-5.6 Sol by sevenfold.

#Benchmark #China #Models #OpenWeights #Step5 #Stepfun

StepFun opened the Step 5 Preview API on Sept 20: 600B parameters, 1M context, vision input and a $1 per million token price. Weights arrive Oct 15.

#AiModels #China #Moe #OpenWeights #Step5 #Stepfun