Did Codex Reset
GitHub

StepFun Step 5 Preview scores 44 on Artificial Analysis Intelligence Index at $0.71 per task

StepFun

StepFun quoted an Artificial Analysis evaluation of Step 5 Preview, saying the model scored 44 on the Artificial Analysis Intelligence Index at $0.71 per task and that more is coming on Oct 15.

According to Artificial Analysis, Step 5 Preview is StepFun’s new flagship: a 600B total / 27B active MoE model with a 1M-token context window, text, image and video input, and text output. It listed pricing at $1 / $2.70 per 1M input/output tokens (cached input $0.05/M), availability on StepFun’s first-party API, and open weights planned for October 15. Artificial Analysis put cost at about $0.72 per Intelligence Index task, matching Kimi K3 (max) at score 44 and about 2.8x lower cost per task.

The evaluation said frontier reasoning was the largest jump from Step 3.7 Flash, including 46% on Humanity’s Last Exam and 21% on CritPt. It also reported 42% AA-Omniscience accuracy with more hallucination than GLM-5.3, and said agentic scores lagged peers at a similar Intelligence Index level, including 1,566 Elo on GDPval-AA.