SchemaBridge Description - RSchema BERTScore F1 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 bERTScore F1 (native). Higher is better.

Metric: BERTScore F1 (native). Source: arxiv.org. Saturation forecast: Around July 2027. 5 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1Llama 3.1 8B Instruct0.863#4545
2CodeLlama-13B-Instruct-hf0.84#9043
3deepseek-coder-6.7B-instruct0.829#7426
4Claude Sonnet 4.60.817#165
5GPT-5.4 Mini0.816#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-rschema-bertscore-f1-detailed · How It Works · Data refreshed daily, snapshot 2026-11-16.