List your company

Spreadsheet Manipulation and Reasoning on SpreadsheetBench

28.8Soft AccuracyTabClaw
14.2418.0221.825.58Jun 9, 2026

Evaluation results

MethodLinks
TabClawCategory=Ours, Backbone=DeepSeek-v3.22026.0628.824.5
BinderCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0624.118.9
SpreadsheetLLMCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0623.319.8
SheetAgentCategory=Autonomous Agent, Backbone=DeepSeek-v3.22026.0622.318.5
VanillaCategory=Direct Inference, Backbone=DeepSeek-v3.22026.0618.815.5
ReActCategory=Autonomous Agent, Backbone=DeepSeek-v3.22026.0618.816.1
Chain-of-TableCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0614.812.9
NormTabCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0614.812.2