Anthropic · frontier
Is Claude Opus 4.7 good?
Claude Opus 4.7 places top-3 in 1 of 6 task areas we tested. It's strongest at chef / home cooking and weakest at landing pages. It's a top-tier (premium) option ($30/1M tokens).
Our take
Strengths
Weak spots
How Claude Opus 4.7 ranks, task by task
Rank is the position in each field; Rating is the absolute quality bar. A model can be rated "Excellent" yet sit mid-table when the whole field is strong — and vice-versa.
| Task area | Rank | Percentile | Rating |
|---|---|---|---|
| Chef / Home Cooking | #2 of 50 |
|
Strong |
| Investor & Pitch | #9 of 50 |
|
Strong |
| AI Strategy | #10 of 50 |
|
Strong |
| Content & Brand | #19 of 50 |
|
Strong |
| Executive Assistant | #21 of 50 |
|
Strong |
| Landing Pages | #30 of 56 |
|
Strong |
Industry benchmarks
Standardized third-party scores, shown for context — these are independent of our real-task tests.
Artificial Analysis indices
Design Arena — human preference (avg Elo 1307)
Headline vs. real-world
Claude Opus 4.7 looks strong on paper but underperforms on our real-world tasks relative to its Intelligence Index.
Source: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings). (2026-06-29) · Source: Design Arena (www.designarena.ai) via OpenRouter (openrouter.ai/rankings). (2026-06-29)
Like Claude Opus 4.7, but…
Similar
- Claude Opus 4.6 Frontier
- GPT-5.4 Premium
- Claude Sonnet 4.6 Premium
Cheaper
- Qwen3.7 Max 83% cheaper
- Gemini 3.5 Flash 65% cheaper
- Gemini 3.1 Pro Preview 53% cheaper
Smarter
- GPT-5.5 +12 pts
- Gemini 3.1 Pro Preview +12 pts
- Claude Opus 4.8 +7 pts
Frequently asked
Is Claude Opus 4.7 any good?
Claude Opus 4.7 places top-3 in 1 of 6 task areas we benchmarked. It performs best at Chef / Home Cooking.
What is Claude Opus 4.7 best at?
Its strongest task areas are Chef / Home Cooking, Investor & Pitch, AI Strategy.
Is there a cheaper alternative to Claude Opus 4.7?
Qwen3.7 Max is 83% cheaper and stays competitive on quality across our benchmarks.
How we test & rank
Each model is scored on real tasks across 6 task areas by an LLM judge with deterministic checks. "Percentile" is this model's standing within each area's field; the cross-area figures are rank-based so quality and Elo scores are never mixed.
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