Maternity Case Record Digitisation - Free-Text WER: leaderboard
A 2026 benchmark of multimodal LLMs for digitising a handwritten South African maternity case record: each form image is extracted into a JSON template and compared field by field with reference JSON files. This board is the word error rate on the free-text fields of the real registry forms, times 100 and able to exceed 100; lower is better.
Metric: Word error rate on the free-text fields (from 0 upward, times 100; can exceed 100), on the real redacted Maternity Case Records (49 forms for Gemini and OpenAI models; 46 for Claude) of the UBOMI BUHLE Pregnancy Exposure Registry (South Africa), one form image per call with the optimised extraction prompt and JSON template, temperature 0, two attempts at valid JSON; lower is better. Source: arxiv.org. Saturation forecast: Around September 2027. 9 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 50 | #99 |
| 2 | Gemini 3.1 Flash Lite | 53 | #371 |
| 3 | Claude Sonnet 4.6 | 56 | #173 |
| 4 | GPT-5.4 | 56 | #135 |
| 5 | Gemini 2.5 Pro | 60 | #276 |
| 6 | Gemini 2.5 Flash | 60 | #463 |
| 7 | Claude Opus 4.6 | 72 | #113 |
| 8 | GPT-5 Mini | 72 | #364 |
| 9 | Claude Haiku 4.5 | 82 | #583 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=maternity-case-record-digitisation-free-text-wer · How It Works · Data refreshed daily, snapshot 2026-11-21.