ESQ-Bench (Tier 1): leaderboard
ESQ-Bench (2026) is an Oracle-first enterprise NL2SQL benchmark with three schema-complexity tiers over 550 gold-validated question-query pairs. Execution match is the share of executed queries whose result set matches the reference on the Tier 1 schemas under schema-linked prompting, in percent, higher is better.
Metric: Execution match (%; among executed queries, the result set matches the reference; Tier-1 schemas; over ESQ-Bench's Oracle-first enterprise NL2SQL question bank; execution match among executed queries; schema-linked prompt, temperature 0). Source: arxiv.org. Saturation forecast: Around February 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 87.4 | #158 |
| 2 | GPT-4o | 79.8 | #741 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=esq-bench-tier-1 · How It Works · Data refreshed daily, snapshot 2026-11-12.