Structured Output Benchmark — leaderboard
Structured-output benchmark measuring schema-constrained generation with value accuracy, faithfulness, JSON validity, path recall, type safety, and perfect-output rates.
Metric: Overall (%). Source: interfaze.ai. Status: saturated. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 87 |
| 2 | Gemini 3.1 Pro (Preview) | 86.9 |
| 3 | GLM-5.1 | 86.6 |
| 4 | Claude Opus 4.7 | 86.4 |
| 5 | Claude Sonnet 5 | 86.2 |
| 6 | GLM-4.7 | 86.1 |
| 7 | Qwen 3.5 35B A3B | 86.1 |
| 8 | GPT-5.5 | 86 |
| 9 | Gemini 2.5 Flash | 86 |
| 10 | Qwen 3 235B A22B | 85.7 |
| 11 | Claude Sonnet 4.6 | 85.4 |
| 12 | Claude Opus 4.6 | 85.3 |
| 13 | Kimi K2.6 | 85.3 |
| 14 | DeepSeek V4 Pro | 85.3 |
| 15 | GPT-4.1 | 85 |
Interactive version: theaggregate.ai/benchmark?slug=structured-output-benchmark · How the rankings work · Data refreshed daily, snapshot 2026-07-22.