Step 3.7 Flash — benchmark results

StepFun's speed-oriented Step 3.7 Flash model. Provider: StepFun. Released 2026-05-29. Access: API.

Unified ELO 1751 ± 20, rank #196 of 1776 rated models, from 70 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA TAU-2 Bench98.54Accuracy (%)99
CLAW-Eval80.5Score (%)92
OpenClawProBench72Overall Score (%)90.9
CritPt2.3Accuracy (self-reported)83.5
AA Omniscience - Software Engineering (SWE) - Rust68Accuracy (%)83.1
AA Terminal-Bench Hard35.61Accuracy (%)82
AA IFBench67.28Accuracy (%)80.4
AA Humanity's Last Exam19.93Accuracy (%)80.1
AA CritPt2.29Accuracy (%)78.4
AA Long Context Reasoning63.67Accuracy (%)78.2
AA Omniscience - Science, Engineering & Mathematics35.5Accuracy (%)77.9
AA MMMU-Pro75.32Accuracy (%)77.6

Interactive version: theaggregate.ai/model?slug=step-3-7-flash · How the rankings work · Data refreshed daily, snapshot 2026-07-22.