OfficeQA Pro

A benchmark for grounded reasoning over office-style documents, spreadsheets, charts, and business artifacts.

13 models

Higher is better.

1Claude Opus 5.5Anthropic67.7
2Claude Opus 5Anthropic66.9
3Claude Opus 4.8Anthropic66.2
4Hy4 previewTencent66.2
5Claude Sonnet 5.5Anthropic65.6
6Kimi K3Moonshot AI63.3
7GLM-5.3-FlashZ.AI62.4
8Step 5 PreviewStepFun60.3
9Claude Fable 5Anthropic57.9
10GPT-5.5OpenAI54.1
11GPT-5.4OpenAI53.2
12MiniMax M3MiniMax45.1
13Claude Opus 4.7 (Adaptive)Anthropic43.6

Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric: Grounded QA over office artifacts
Tasks: Document and spreadsheet tasks
Authors: OfficeQA Pro authors (2026)
Source: OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning ↗

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.