SchemaBridge Direct Gold Text - RSchema Tables F1: leaderboard
SchemaBridge evaluates relational-schema design, description and reconstruction. Experiment-specific cohorts retain their own populations. Graph edit distance, hallucination, structural F1, text similarity and identifier coverage are distinct measurements. Round-trip runs use a fixed GPT-4o Mini constructor with each named describer. This board reports micro structural Tables F1 (0-1). Higher is better.
Metric: Micro structural Tables F1 (0-1). Source: arxiv.org. Saturation forecast: Around 2029. 8 models tracked. Updated 14 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 0.602 | #165 | |
| 2 | GPT-5.4 Mini | 0.6 | #398 | |
| 3 | deepseek-coder-6.7b-instruct | SchemaBridge run: Few-shot | 0.6 | |
| 4 | CodeLlama-13b-Instruct-hf | SchemaBridge run: Few-shot | 0.6 | |
| 5 | Meta-Llama-3.1-8B-Instruct | SchemaBridge run: Few-shot | 0.585 | |
| 6 | Llama 3.1 8B Instruct | 0.572 | #4545 | |
| 7 | deepseek-coder-6.7B-instruct | 0.566 | #7426 | |
| 8 | CodeLlama-13B-Instruct-hf | 0.554 | #9043 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=schemabridge-direct-gold-text-rschema-tables-f1 · How It Works · Data refreshed daily, snapshot 2026-11-16.