SpreadsheetBench v2 - Debugging Accuracy: leaderboard
Metric: Task accuracy (%). Source: spreadsheetbench.github.io. Saturation forecast: Around 2030. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 12 |
| 2 | DeepSeek V3.2 | 10 |
| 3 | GPT-5.2 | 8 |
| 4 | Gemini 3.1 Pro (Preview) | 7 |
| 5 | GLM-5 | 7 |
| 6 | Kimi K2.5 | 4 |
| 7 | MiniMax-M2.5 | 4 |
| 8 | Qwen 3.5 397B A17B | 3 |
Interactive version: theaggregate.ai/benchmark?slug=spreadsheetbench-v2-debugging-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-02.