Back to Leaderboard

Claude Opus 4.6

Anthropic
Anthropic200K context$40.75/1K pages2026-02-15
Overall Rank
#8
of 27 models
Overall Score
78.8
avg across benchmarks
Best Task
Key Information Extraction
89.8
Weakest Task
Visual QA
64.4

Benchmark Performance

OlmOCR Benchv1.0
12/27
OverallArXiv MathH&FLong/TinyMulti-ColOld ScansScans MathTables
69.386.635.573.879.520.583.084.5
OmniDocBenchv1.5
7/27
OverallText Edit↓CDM↑TEDS↑TEDS-S↑Read Order↓
85.90.15188.584.489.10.136
IDP Core Benchv1.0
5/27
OverallKIEOCRTableVQA
81.189.874.096.064.4

Capability Profile

Strength Analysis

Auto-generated from benchmark scores

Strengths

  • Key Information Extraction89.8
  • Formula88.5

Weaknesses

  • Visual QA64.4
  • Layout & Order83.0