Anthropic · 9 models tested
Which Anthropic model should you use?
Anthropic's strongest model is Claude Opus 4.8 (avg percentile 87, top-3 in 7 of 22 task areas). On a budget, Claude Sonnet 4.6 stays close at $18/1M.
The Anthropic lineup, ranked
| Model | Overall | Code & data | Writing & comms | Business & strategy | Creative & visual | AA Intel† | Top-3s | Price / 1M |
|---|---|---|---|---|---|---|---|---|
|
Claude Opus 4.8
Best overall
22 task areas
|
87 | 84★ | 87★ | 90★ | 82 | 56 | 7 | $30.0 |
|
Claude Opus 4.7
6 task areas
|
84 | — | 90 | 91 | 69 | 54 | 1 | $30.0 |
|
Claude Opus 4.5
22 task areas
|
79 | 64 | 81 | 78 | 92★ | — | 3 | $30.0 |
|
Claude Sonnet 4.6
Best value
22 task areas
|
76 | 57 | 72 | 85 | 83 | 47 | 4 | $18.0 |
|
Claude Opus 4.6
22 task areas
|
73 | 59 | 71 | 82 | 73 | — | 2 | $30.0 |
|
Claude Fable 5
22 task areas
|
68 | 40 | 73 | 73 | 81 | 60 | 6 | $60.0 |
|
Claude Sonnet 4.5
22 task areas
|
61 | 55 | 75 | 53 | 62 | 36 | 1 | $18.0 |
|
Claude Sonnet 5
22 task areas
|
60 | 64 | 56 | 63 | 56 | 53 | 1 | $12.0 |
|
Claude Haiku 4.5
22 task areas
|
39 | 29 | 44 | 42 | 34 | 30 | 0 | $6.0 |
Area figures are mean percentile ranks; 85+ 70–84 55–69 40–54 <40 · ★ = the lineup's best in that area. †AA Intel = Artificial Analysis Intelligence Index, third-party reference only.
Best Anthropic model, task by task
The lineup's strongest model in each task area — and where it lands against every model we test, not just Anthropic's. #1 means it beats every rival lab too.
Where another lab clearly wins
Honesty corner: the task areas where even Anthropic's best model ranks furthest from the top. See who actually leads there.
Frequently asked
What is the best Anthropic model?
Claude Opus 4.8 — the highest average percentile (87) across the task areas we benchmarked, with 7 top-3 finishes.
What is the best cheap Anthropic model?
Claude Sonnet 4.6: within 10 percentile points of Claude Opus 4.8 at $18 per 1M tokens.
Where does Anthropic lead every other lab?
A Anthropic model holds the outright #1 spot in 8 task areas: AI Strategy, Creative & Comedy, Executive Assistant, Investor & Pitch, Legal & HR, Sales, Summarization & Meeting Notes, Training & Education.
This page is Spring Prompt, running
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