step-1v-32k: benchmark results

Provider: StepFun. Access: API.

Aggregate Rating 1478 ± 48, rank #2677 of 10489 rated models, from 15 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
CoreCognition Image only Permanence87.5Accuracy (%)96.7
CoreCognition Image only Hierarchy68.04Accuracy (%)96.1
CoreCognition Image only Mean64.5Accuracy (%)90.8
CoreCognition Image only Mechanicalreasoning60.32Accuracy (%)90.8
CoreCognition Image only Intentionality77.06Accuracy (%)88.2
CoreCognition Image only Spatiality69.13Accuracy (%)88.2
CoreCognition Image only Toolusing98.96Accuracy (%)82.2
CoreCognition Image only Boundary76.5Accuracy (%)78.9
CoreCognition Image only Perceptualconstancy65.5Accuracy (%)76.3
CoreCognition Image only Intuitivephysics56.03Accuracy (%)60.5
CoreCognition Image only Continuity61.33Accuracy (%)56.6
CoreCognition Image only Perspective33.05Accuracy (%)41.4

Interactive version: theaggregate.ai/model?slug=step-1v-32k · How It Works · Data refreshed daily, snapshot 2026-11-20.