All AutosheetBench tasks

Chart sorted values, source unchanged

ChartsMediumLLM judge scoringLatest run

Passed by 9 of 12 products on their best model. Failed: Copilot in Excel, Gemini in Sheets and ChatGPT in Sheets.

Prompt and starting spreadsheet

Task files

Prompt

Create a chart sorted from highest to lowest without modifying the source table.

Data

Starting sheet
AB
1ProductRevenue
2Alpha8100
3Beta12700
4Gamma9400
5Delta15300
6Epsilon11500
7Zeta6900
8Eta14100
9Theta10200
10Iota13300
11Kappa7600

Report

Starting sheet
A
1

Expected result

How each AI configuration did

17 of 22 passRun files
Excel and Sheets · All types · All models · All outputs

Pass criteria

All must be met, checked by an LLM judge.
  • The Data sheet is unchanged
  • A helper table sorted by revenue, highest first
  • A bar chart from that table, highest first

Passed 17

Autosheet in Excel Opus 5.5 · Medium

Passed26s · 8¢

Autosheet in Excel GPT-6 Astra · Low

Passed15s · 12¢

Autosheet in Sheets Opus 5.5 · Medium

Passed17s · 6¢

Autosheet in Excel Sonnet 5.5 · Medium

Passed9s · 2¢

Claude in Excel Opus 5.5 · High

Passed38s · 14¢

Autosheet in Excel Haiku 5.5 · High

Passed18s · 0.3¢

Claude in Sheets Opus 5.5 · Medium

Passed23s · 6¢

Claude Code + Sheets MCP Opus 5.5 · Medium

Passed44s · N/A

Autosheet in Excel GPT-6.1 Sol · Medium

Passed42s · 7¢

Claude Code + Autosheet MCP Opus 5.5 · Medium

Passed1m 7s · N/A

ChatGPT + Autosheet MCP GPT-6 · Medium

Passed48s · N/A

ChatGPT in Excel GPT-6 Astra · Medium

Passed31s · 58¢

Claude in Excel Haiku 5.5 · High

Passed20s · 0.7¢

Claude in Sheets Haiku 5.5 · High

Passed21s · 0.6¢

Autosheet in Excel GPT-6 Luna · High

Passed52s · 1¢

ChatGPT + Drive MCP GPT-6 · Medium

Passed35s · N/A

ChatGPT in Excel GPT-6 Luna · High

Passed58s · 0.8¢

Failed 5

Autosheet in Sheets Haiku 5.5 · High

Failed52s · 0.5¢

Why it failed: The Data sheet was modified: 22 cells were added across Data!D1:E11 and the chart was placed on the Data sheet.

Copilot in Excel Auto

Failed44s · N/A

Why it failed: The highest bar (Delta) appears at the bottom and the lowest bar (Zeta) appears at the top instead of having the highest bar at the top.

Gemini in Sheets Auto

Failed26s · N/A

Why it failed: The helper table was arranged with products in alphabetical order across rows rather than sorted from highest revenue to lowest revenue.

ChatGPT in Sheets GPT-6 Astra · Medium

Failed1m 16s · 67¢

Why it failed: No native column or bar chart was created.

ChatGPT in Sheets GPT-6 Luna · High

Failed1m 2s · 0.6¢

Why it failed: No chart was created.