Maternity Case Record Digitisation - Discrete Fields Macro F1: 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 macro-averaged F1 on a 0 to 100 scale over the discrete selection fields of the real registry forms; higher is better.

Metric: Macro-averaged F1 (from 0-1, times 100) over the discrete selection fields treated as multi-class classification with null as its own class, 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; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3.1 Pro (Preview)66#99
2GPT-5.456#135
3Gemini 3.1 Flash Lite56#371
4Gemini 2.5 Flash51#463
5Gemini 2.5 Pro46#276
6Claude Opus 4.646#113
7Claude Sonnet 4.645#173
8Claude Haiku 4.534#583
9GPT-5 Mini15#364

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-discrete-fields-macro-f1 · How It Works · Data refreshed daily, snapshot 2026-11-21.