FinED-Bench: leaderboard

Metric: F1 (%; harmonic mean of pooled precision and recall over all injected errors in 973 Chinese financial documents from 2025, nine real-world scenarios; a predicted error counts as correct when the extracted sentence matches or contains the annotated sentence and the error type is classified correctly; the paper's error-detection prompt over each document, reasoning models with thinking on unless marked no thinking). Source: arxiv.org. Saturation forecast: Estimated already saturated. 11 models tracked.

Top models

#ModelScore
1GPT-4o48.34
2Qwen 3 14B43.15
3Qwen 3 8B39.99
4Qwen 3 14B (Non-reasoning)27.19
5DeepSeek R1 0528 Qwen3 8B23.31
6GPT-4o Mini18.9
7Qwen 3 8B (Non-reasoning)17.44
8Qwen 2.5 7B Instruct9.85
9GPT-3.5 Turbo7.17

Interactive version: theaggregate.ai/benchmark?slug=fined-bench · How It Works · Data refreshed daily, snapshot 2026-09-26.