Back to Leaderboard

GPT-5.4

OpenAI
OpenAI400K context2026-03-05
Overall Rank
#4
of 18 models
Overall Score
81.0
avg across benchmarks
Best Task
Text Extraction
91.1
Weakest Task
Visual QA
78.2

Benchmark Performance

OlmOCR Benchv1.0
7/18
OverallMathTablePresentAbsentOrder
73.483.191.166.925.274.7
OmniDocBenchv1.5
9/18
OverallText Edit↓CDM↑TEDS↑TEDS-S↑Read Order↓
85.30.08983.481.386.70.077
IDP Core Benchv1.0
2/18
OverallKIEOCRTableVQA
84.485.769.194.878.2

Capability Profile

Strength Analysis

Auto-generated from benchmark scores

Strengths

  • Text Extraction91.1
  • Table Understanding89.1

Weaknesses

  • Visual QA78.2
  • Formula83.4