priorsBuilder

OpenAI

GPT-5.3

Measured on 5 benchmarks.

General frontier

42.2

#51 of 136 models · best at medium

includes estimates: 5% of the mix is measured for this model

Your mix’s benchmarks

1 of 8 measured

reasoning

BullshitBench v2

196 results
  1. #96medium of 19640.0%

Leader: Opus 4.8 · 95.0%

All 5 benchmarks

show 4 more

coding

SWE Atlas-QnA

22 results
  1. #14xhigh of 2232.6%

Leader: Opus 5 · 63.2%

composite

Arena Text

220 results
  1. #67xhigh of 2201450

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #56xhigh of 2201467

Leader: Gemini 4 · 1553

other

T3 Code usage share

85 results
  1. #84xhigh of 850.0%

Leader: Opus 5.5 · 29.1%