WildHandBench - Free Text: leaderboard
The free-text part of WildHandBench (2026): transcription of real-world handwritten text documents, scored by normalized edit distance to the ground truth in percent; lower is better.
Metric: Normalized edit distance (%; lower is better; handwritten free-text transcription over the text documents of 500 real handwritten documents (free text, tables and formulas) in four languages and nine real-world scenarios such as notes, forms, whiteboards and exam sheets). Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 24.32 | #94 |
| 2 | Kimi K3 | 26.07 | #54 |
| 3 | Gemini 3.5 Flash | 27.1 | #101 |
| 4 | Qwen 3 VL 235B A22B (Thinking) | 34.3 | #451 |
| 5 | Qwen 3.5 Plus | 34.8 | #232 |
| 6 | Claude Opus 4.8 | 53.3 | #56 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=wildhandbench-free-text · How It Works · Data refreshed daily, snapshot 2026-11-13.