All AutosheetBench tasks

Show numbers as K and M, negatives in red

FormattingEasyLLM judge scoringLatest run

Passed by 11 of 12 products on their best model. Failed: Gemini in Sheets.

Prompt and starting spreadsheet

Task files

Prompt

Format the values so that: - 1,200,000 displays like 1.2M - 15,000 displays like 15K - negative values display in red with parentheses

Starting spreadsheet

Data
A
1Value
21200000
31500000
415000
522000
6-5400
7-12500

Expected result

How each AI configuration did

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

Pass criteria

All must be met, checked by an LLM judge.
  • Millions show as 1.2M
  • Thousands show as 15K
  • Negatives show in red, in parentheses
  • The values stay numbers

Passed 21

Autosheet in Excel Opus 5.5 · Medium

Passed23s · 6¢

Autosheet in Excel GPT-6 Astra · Low

Passed10s · 6¢

Autosheet in Sheets Opus 5.5 · Medium

Passed37s · 9¢

Autosheet in Excel Sonnet 5.5 · Medium

Passed11s · 2¢

Claude in Excel Opus 5.5 · High

Passed47s · 18¢

Autosheet in Excel Haiku 5.5 · High

Passed38s · 0.5¢

Claude in Sheets Opus 5.5 · Medium

Passed38s · 15¢

Claude Code + Sheets MCP Opus 5.5 · Medium

Passed1m 2s · N/A

Autosheet in Excel GPT-6.1 Sol · Medium

Passed17s · 6¢

Claude Code + Autosheet MCP Opus 5.5 · Medium

Passed1m 41s · N/A

ChatGPT + Autosheet MCP GPT-6 · Medium

Passed1m 10s · N/A

Autosheet in Sheets Haiku 5.5 · High

Passed1m 44s · 1¢

ChatGPT in Excel GPT-6 Astra · Medium

Passed30s · 41¢

Claude in Excel Haiku 5.5 · High

Passed51s · 1¢

Claude in Sheets Haiku 5.5 · High

Passed1m 47s · 3¢

Autosheet in Excel GPT-6 Luna · High

Passed1m 22s · 2¢

Copilot in Excel Auto

Passed27s · N/A

ChatGPT + Drive MCP GPT-6 · Medium

Passed53s · N/A

ChatGPT in Excel GPT-6 Luna · High

Passed1m 26s · 0.9¢

ChatGPT in Sheets GPT-6 Astra · Medium

Passed33s · 33¢

ChatGPT in Sheets GPT-6 Luna · High

Passed56s · 0.6¢

Failed 1

Gemini in Sheets Auto

Failed5m · N/A

Why it failed: Still working at the 5-minute time limit. Cells A2:A3 display '#VALUE!' errors in Document 1 rather than displaying formatted values like 1.2M.