SpreadsheetBench v2 - Financial Modeling Cell Match: leaderboard

Metric: Cell-match score (%, Modif.) on the 100 financial-modeling tasks under the SWE-agent-based scaffold (bash, view_xlsx and submit tools, at most 50 turns), thinking mode on, high reasoning effort for GPT-5.2 and Gemini 3.1 Pro. Source: arxiv.org. Saturation forecast: Around December 2026. 8 models tracked.

Top models

#ModelScore
1Claude Opus 4.6 (Thinking)89.69
2GPT-5.2 (High)87.07
3Gemini 3.1 Pro (Preview)85.26
4GLM-578.44
5Qwen 3.5 397B A17B69.59
6Kimi K2.564.86
7MiniMax-M2.562.05
8DeepSeek V3.2 (Thinking)59.78

Interactive version: theaggregate.ai/benchmark?slug=spreadsheetbench-v2-financial-modeling-cell-match · How It Works · Data refreshed daily, snapshot 2026-09-29.