Is claude-opus-4.5-high good at Structured Output?
claude-opus-4.5-high ranks #25 of 43 for Structured Output — strong. The top pick for this task is qwen3.7-max-low.
claude-opus-4.5-high on each Structured Output sub-task
| Extraction | 91.4/100 | #17 |
| Missing & Ambiguous Data | 90.0/100 | #32 |
| Schema Adherence | 86.4/100 | #29 |
| Transformation | 83.6/100 | #31 |
| Noisy Structured Output Test | 79.8/100 | #34 |
Real examples, graded
WeakKey order requirement 90/100
“The model successfully extracted all required data and adhered perfectly to the strict key ordering requirement. However, it failed the negative constraint to return *only* the JSON object by including markdown fences, which prevents the raw response from being parsed directly.”
WeakEvent details to typed schema 90/100
“The model perfectly extracted all values, adhered to the schema types, and correctly used null for the missing date. However, it failed the negative constraint by including markdown fences instead of returning raw, directly parseable JSON.”
WeakArray deduplication 6/100
“The model failed to follow instructions by including markdown fences, which resulted in invalid JSON and wrapper text violations.”
Frequently asked
Is claude-opus-4.5-high good at Structured Output?
claude-opus-4.5-high ranks #25 of 43 models we tested for Structured Output, scoring strong.
What is claude-opus-4.5-high's strongest Structured Output skill?
Its best sub-task here is Extraction.
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