Skip to content

Benchmark

SpreadsheetBench 2 Template

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
7
Scores recorded
7
Sources
1
IQ weight
0.98

Score distribution

26.5%
Mean
25.8%
Median
13.5%
Std dev
17.5%
10th percentile
35.1%
90th percentile
7.2%
Min
7.2%52.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.98 for chat models, measured over 7 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 4.6scaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data52.6%
2GPT 5.2scaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data35.1%
3Gemini 3.1 Pro Previewscaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data28.9%
4DeepSeek V3.2scaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data25.8%
5Kimi K2.5scaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data18.6%
6Qwen3.5 397B A17Bscaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data17.5%
7MiniMax M2.5scaffold: swe-agent, date: 2026-03-26· SpreadsheetBench 2 leaderboard data7.2%

Data credits: SpreadsheetBench (Renmin University of China and contributors), spreadsheetbench.github.io.