All AutosheetBench tasks

Highlight whole rows above a threshold

FormattingMediumDeterministic scoringLatest run

Passed by 10 of 12 products on their best model. Failed: Copilot in Excel and ChatGPT in Sheets.

Prompt and starting spreadsheet

Task files

Prompt

Highlight entire rows where revenue is greater than 10000 with a light green fill.

Starting spreadsheet

Data
ABC
1DateProductRevenue
22025-03-01Alpha8400
32025-03-02Beta11200
42025-03-03Gamma9700
52025-03-04Delta13800
62025-03-05Epsilon7600
72025-03-06Zeta15100
82025-03-07Eta8900
92025-03-08Theta12800

Expected result

How each AI configuration did

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

Pass criteria

All must be met, checked automatically.
  • A conditional rule =$C2>10000 highlights A2:C9 in light green

Passed 19

Autosheet in Excel Opus 5.5 · Medium

Passed7s · 2¢

Autosheet in Excel GPT-6 Astra · Low

Passed5s · 4¢

Autosheet in Sheets Opus 5.5 · Medium

Passed7s · 2¢

Autosheet in Excel Sonnet 5.5 · Medium

Passed4s · 0.9¢

Claude in Excel Opus 5.5 · High

Passed18s · 11¢

Autosheet in Excel Haiku 5.5 · High

Passed9s · 0.1¢

Claude in Sheets Opus 5.5 · Medium

Passed20s · 4¢

Claude Code + Sheets MCP Opus 5.5 · Medium

Passed32s · N/A

Autosheet in Excel GPT-6.1 Sol · Medium

Passed9s · 1¢

Claude Code + Autosheet MCP Opus 5.5 · Medium

Passed31s · N/A

ChatGPT + Autosheet MCP GPT-6 · Medium

Passed33s · N/A

Autosheet in Sheets Haiku 5.5 · High

Passed17s · 0.1¢

ChatGPT in Excel GPT-6 Astra · Medium

Passed13s · 37¢

Claude in Excel Haiku 5.5 · High

Passed7s · 0.4¢

Claude in Sheets Haiku 5.5 · High

Passed26s · 0.5¢

Autosheet in Excel GPT-6 Luna · High

Passed7s · 0.1¢

ChatGPT + Drive MCP GPT-6 · Medium

Passed34s · N/A

ChatGPT in Excel GPT-6 Luna · High

Passed39s · 0.6¢

Gemini in Sheets Auto

Passed36s · N/A

Failed 3

Copilot in Excel Auto

Failed19s · N/A

Why it failed: No conditional rule on A2:C9.

ChatGPT in Sheets GPT-6 Astra · Medium

Failed27s · 33¢

Why it failed: No conditional rule on A2:C9.

ChatGPT in Sheets GPT-6 Luna · High

Failed17s · 0.3¢

Why it failed: No conditional rule on A2:C9.