PulseAugur
中
实时 03:47:04
English(EN) Text-to-SQL on Oracle: two prompt rules took a light model from 35% to 64%

Gemini 2.5 Flash 在 Oracle Text-to-SQL 基准测试中击败 Claude Sonnet 4.5

一个名为 UniQL 的新基准测试旨在将自然语言问题转换为各种方言的 SQL 查询,突显了大型语言模型的性能差异。在 UniQL 的 Oracle 方言上,Claude Sonnet 4.5 的准确率为 49.61%,而 Gemini 2.5 Pro 的准确率为 26.14%。通过实施针对 Oracle 语法和保留字量身定制的特定提示规则,Gemini 2.5 Flash 模型能够将其单次尝试的准确率从 35% 提高到 64.5%,超过了之前报告的最佳分数。 AI

影响 展示了提示工程和特定数据库规则对 LLM 在 Text-to-SQL 任务中性能的显著影响。

排序理由 新基准测试和模型在特定任务上的性能比较。[lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

Gemini 2.5 Flash 在 Oracle Text-to-SQL 基准测试中击败 Claude Sonnet 4.5

本文如何被排名

Signal score
48 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
新基准测试和模型在特定任务上的性能比较。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Ashish sinha ·

    Oracle 上的 Text-to-SQL:两条提示规则将轻量级模型从 35% 提升至 64%

    <p><a href="https://github.com/JerryGao818/UniQL" rel="noopener noreferrer">UniQL</a> is a new benchmark that translates BIRD's 1,534 text-to-SQL questions into 16 SQL dialects. Its Oracle column is hard. The best published result is <strong>49.61%</strong> (Claude Sonnet 4.5), a…