A new benchmark called GDP.xlsx has been developed to evaluate the capabilities of frontier AI models in handling realistic workplace tasks. The benchmark consists of 70 requests from 13 professional domains, each accompanied by an expert rubric and the relevant workbook or CSV file. Across 20 model configurations, including leading models from Google, Anthropic, and OpenAI, the top performer, Gemini 4 Argon, achieved a pass rate of only 38.3%, indicating significant room for improvement in AI's ability to manage complex professional workbooks. AI
IMPACT Highlights significant limitations in current frontier models for complex professional tasks, indicating a need for improved agentic capabilities in spreadsheet and data manipulation.
RANK_REASON The cluster contains a research paper introducing a new benchmark for evaluating AI models. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →