Anthropic
VS
Google
Anthropic vs Google: which AI lab wins at real work?
22 task areas · each lab's best model per task · same graded test runs · rank comparison only.
Anthropic's best model out-ranks Google's in 18 of 22 task areas we tested; Google takes 4. Lineup prices run $6–$60/1M for Anthropic vs $0.75–$14/1M for Google.
Task by task — each lab's best
For every task area, we take each lab's highest-ranked model and compare where they land against every model we test.
| Task area | Anthropic's best | Google's best | Winner |
|---|---|---|---|
| Structured Output |
#61 / 113
|
#15 / 113
|
|
| Research & Competitive Analysis |
#3 / 110
|
#42 / 110
|
Anthropic |
| Data & Analytics |
#2 / 110
|
#34 / 110
|
Anthropic |
| Sales |
#1 / 110
|
#32 / 110
|
Anthropic |
| Product & Project Management |
#2 / 110
|
#25 / 110
|
Anthropic |
| AI Strategy |
#1 / 126
|
#23 / 126
|
Anthropic |
| Coding |
#3 / 115
|
#22 / 115
|
Anthropic |
| Knowledge & Docs |
#2 / 110
|
#19 / 110
|
Anthropic |
| Legal & HR |
#1 / 110
|
#17 / 110
|
Anthropic |
| Executive Assistant |
#1 / 112
|
#14 / 112
|
Anthropic |
| Landing Pages |
#1 / 72
|
#14 / 72
|
Anthropic |
| Translation & Localization |
#2 / 110
|
#14 / 110
|
Anthropic |
| Content & Brand |
#1 / 124
|
#11 / 124
|
Anthropic |
| Investor & Pitch |
#3 / 66
|
#13 / 66
|
Anthropic |
| RAG, Safety & Grounding |
#11 / 113
|
#5 / 113
|
|
| Chef / Home Cooking |
#3 / 126
|
#6 / 126
|
Anthropic |
| Creative & Comedy |
#1 / 110
|
#4 / 110
|
Anthropic |
| Presentations & Decks |
#20 / 110
|
#17 / 110
|
|
| Summarization & Meeting Notes |
#1 / 110
|
#4 / 110
|
Anthropic |
| Training & Education |
#1 / 110
|
#4 / 110
|
Anthropic |
| Customer Support |
#1 / 113
|
#2 / 113
|
Anthropic |
| Frontend & Landing Pages |
#2 / 109
|
#1 / 109
|
Rank = the model's position among every model config we tested in that task area (lower is better). Sorted by biggest gap first.
Frequently asked
Is Anthropic better than Google for real work?
Comparing each lab's best model per task area across 22 areas, Anthropic ranks higher in 18 and Google in 4.
What is Anthropic's strongest model?
Claude Opus 4.8 — avg percentile 87 across our task areas.
What is Google's strongest model?
Gemini 3.1 Pro Preview — avg percentile 82 across our task areas.
Where does Anthropic beat Google?
Anthropic leads at Research & Competitive Analysis, Data & Analytics, Sales.
Where does Google beat Anthropic?
Google leads at Structured Output, RAG, Safety & Grounding, Presentations & Decks.
More lab match-ups
This page is Spring Prompt, running in public
We just did this for every model. Do it for your prompt.
The rankings above come from running real tasks through real models and scoring every output. Spring Prompt is that same engine — pointed at your prompt, your test cases, and your definition of good.
- Generate test cases from your prompt — no eval set required to start.
- Compare models side by side with quality, cost and latency in one matrix.
- Optimise the winner until the scores say it's ready to ship.
Prompt × model results
12 test cases · 3 evals