Best AI for Excel and spreadsheets
As of 2026-07-18, Anthropic: Claude Opus 4.8 (Fast) is the strongest model for this task in our daily benchmark, where over 350 AI models are evaluated and the best are presented.
Spreadsheets are one of the tasks where AI pays off immediately: you describe what you want and get a formula. But a formula that looks right and returns the wrong number is worse than no formula.
For spreadsheets we rank by instruction-following, because a formula either meets the spec or it does not. Ranked by instruction score, from today’s benchmark.
Today’s ranking
| Model | Provider | Score | Language | Speed | Input price/1M |
|---|---|---|---|---|---|
| Anthropic: Claude Opus 4.8 (Fast) | anthropic | 8.3/10 | 9.3/10 | 131 t/s | $10.00 |
| Anthropic: Claude Haiku 4.5 | anthropic | 8.0/10 | 9.3/10 | 113 t/s | $1.00 |
| Mistral: Mistral Small 3 | mistralai | 7.8/10 | 8.7/10 | 234 t/s | $0.05 |
| Perplexity: Sonar Pro | perplexity | 7.7/10 | 9.0/10 | 68 t/s | $3.00 |
| Cohere: Command R+ (08-2024) | cohere | 7.5/10 | 9.3/10 | 160 t/s | $2.50 |
| Meta: Llama 4 Maverick | meta-llama | 7.2/10 | 8.0/10 | 176 t/s | $0.20 |
| OpenAI: GPT-5.4 Nano | openai | 7.1/10 | 8.0/10 | 157 t/s | $0.20 |
| OpenAI: GPT-5.6 Sol Pro | openai | 6.8/10 | 7.3/10 | 936 t/s | $5.00 |
Describe the sheet, not just the formula
The model cannot see your spreadsheet. Say which columns you have, what is in them and where the result goes: “column A is date, B is amount, C is department — I want the sum per department for the current month in E2”. Without that context the model guesses at cell references, and often guesses wrong.
Test on a few rows first
Always run the formula on a small sample where you know the answer before dragging it down a thousand rows. The common failure is not syntax — the sheet tells you about that — but logic: a condition that misses the edges, or an absolute range that should have been relative.
Frequently asked questions
Other tasks
Doing something else? See which model is best for it.
Read also