Benchmark RadarAI BENCHMARK PROFILE
OmniHandwritingOCR
Evaluates multimodal LLMs and OCR systems on handwritten text and mathematical expression recognition across six subtasks and twelve subsets with 77.57K images.
- Released
- 2026-08-19
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Provides a challenging diagnostic benchmark for realistic handwritten OCR, highlighting failures in complex formulas and visual grounding that existing printed-text benchmarks miss.
Motivation
Multimodal large language models (MLLMs) are increasingly used as OCR systems in document and knowledge-processing pipelines, but their ability to faithfully read real handwriting remains underexplored.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.