priorsBuilder

stepfun · released 2026-09-18

Step 5 Preview

Measured on 15 benchmarks.

General frontier

Not ranked in this mix — too little measured data for its benchmarks.

Your mix’s benchmarks

0 of 8 measured

This model has no results on your mix’s benchmarks.

All 15 benchmarks

show 15 more

agentic

BioMysteryBench

27 results
  1. #17default of 2767.4%

Leader: Sonnet 5.5 · 81.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #21default of 799.2%

Leader: Muse Spark 1.2 · 25.4%

agentic

Legal Research Bench

78 results
  1. #29default of 7840.4%

Leader: Muse Spark 1.3 · 55.3%

agentic

Tax Agent Bench

69 results
  1. #31default of 6927.3%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 4.0 (Vals)

47 results
  1. #16default of 4731.8%

Leader: Opus 5.5 · 65.2%

agentic

Terminal-Bench Science (Vals)

40 results
  1. #31default of 402.9%

Leader: GPT-6 Astra · 62.9%

coding

Code Migration

77 results
  1. #31default of 7739.0%

Leader: Sonnet 5.5 · 69.8%

coding

IOI

44 results
  1. #38default of 4438.1%

Leader: Gemini 4 · 100.0%

coding

Vibe Code Bench v1.1

105 results
  1. #37default of 10570.4%

Leader: Sonnet 5.5 · 92.4%

composite

Vals Index

47 results
  1. #32default of 4749.4%

Leader: Gemini 4 · 68.9%

economics

Excel Modeling Benchmark

74 results
  1. #40default of 7457.1%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #40default of 7850.7%

Leader: Gemini 4 · 65.4%

knowledge

Public Benefits Bench v1.1

50 results
  1. #37default of 5058.1%

Leader: Opus 5 · 76.9%

math

ProofBench v1.1

52 results
  1. #36default of 5242.0%

Leader: Sonnet 5.5 · 100.0%

reasoning

MysteryMechanism

27 results
  1. #24default of 2718.5%

Leader: GPT-6 Astra · 53.2%