Step 1 8K: benchmark results

Provider: StepFun. Access: API.

Aggregate Rating 1514 ± 24, rank #2125 of 10916 rated models, from 141 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
OpenCompass Agent - T-Eval - English (CompassBench 2407)87.2Score (%)88.2
OpenCompass Reasoning - Abductive - English (CompassBench 2405)86Score (%)86.8
OpenCompass Language - Content Summarization - Chinese (CompassBench 2405)25.9Score (%)85.5
OpenCompass Knowledge - Humanities - Chinese (CompassBench 2405)52Score (%)78.9
OpenCompass Language - Sentiment Analysis - English (CompassBench 2405)77.7Score (%)78.9
OpenCompass Reasoning - Inductive - English (CompassBench 2405)34.7Score (%)78.9
OpenCompass LLM - Instruction Following - Chinese (CompassBench 2407)45.5Score (%)76.5
OpenCompass Language - NLP - Chinese (CompassBench 2407)54.3Score (%)76.5
OpenCompass Agent - T-Eval - English (CompassBench 2405)84.4Score (%)73.7
OpenCompass Math - Middle School - English (CompassBench 2405)28.9Score (%)73.7
OpenCompass Language - Sentiment Analysis - Chinese (CompassBench 2405)87.3Score (%)72.4
OpenCompass Agent - T-Eval - Chinese (CompassBench 2405)79.5Score (%)71.1

Interactive version: theaggregate.ai/model?slug=step-1-8k · How It Works · Data refreshed daily, snapshot 2026-11-08.