DocMath · Complong
The more complex long-document numerical reasoning task.
Fino1 · 2025 · published research
Evaluation setup
Author-reported results using the Fino1 evaluation setup. Scores are copied from the repository table, not independently reproduced by Finsider.
Reading this comparison
Higher values indicate better performance under the named study's scoring rules. Published model names are retained. These are historical snapshots, not tests of later versions with similar names.
Boundaries
A benchmark answer score does not establish a model's readiness to deliver a financial diligence engagement. It does not measure CPA sign-off, evidence retention, access controls, or performance on your transaction documents.
Provenance
Transcribed from the linked primary source on September 12, 2026. Evaluation was performed by the source authors, not Finsider. The downloadable CSV includes each value, configuration, and source URL.
Performance Comparison
Reported correctness (%)
DeepSeek-V3
42.33
GPT-4o
39.33
GPT-4.5
39.33
DeepSeek-R1
38.67
GPT-o1-preview
36.67
GPT-o3-mini
35.00
Fino1-14B
27.33
Qwen2.5-72B-Instruct
14.67
Eight selected models from the published comparison, not the complete model roster.
Rank uses displayed scores, and ties share rank. Bars use a 0–100 scale. The source table provides no confidence intervals, so none are drawn.
Source: Fino1 · 2025