SchemaBridge Description - SchemaBench ROUGE L Detailed: 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 rOUGE L (native). Higher is better.

Metric: ROUGE L (native). Source: arxiv.org. Saturation forecast: Around 2037. 5 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1Llama 3.1 8B Instruct0.175#4545
2CodeLlama-13B-Instruct-hf0.145#9043
3deepseek-coder-6.7B-instruct0.141#7426
4Claude Sonnet 4.60.128#165
5GPT-5.4 Mini0.124#398

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=schemabridge-description-schemabench-rouge-l-detailed · How It Works · Data refreshed daily, snapshot 2026-11-16.