paper-with-me

Text-To-SQL 벤치마크

Text-To-SQL on spider

56개 결과 · ⬇ CSV · JSON

Execution Accuracy (Test)

68.32 75.23 82.13 89.04 95.95 2021-09 2026-11 T5-3B + PICARD — 75.1 (2021-09-10) T5-3B + PICARD — 75.1 (2021-09-10) Graphix-3B+PICARD — 77.6 (2023-01-18) Graphix-3B+PICARD — 77.6 (2023-01-18) RESDSQL-3B + NatSQL — 79.9 (2023-02-12) RESDSQL-3B + NatSQL — 79.9 (2023-02-12) DIN-SQL + GPT-4 — 85.3 (2023-04-21) DIN-SQL + GPT-4 — 85.3 (2023-04-21) T5-3B+NatSQL+Token Preprocessing — 78.0 (2023-05-27) T5-3B+NatSQL+Token Preprocessing — 78.0 (2023-05-27) T5-SR — 75.2 (2023-06-14) T5-SR — 75.2 (2023-06-14) C3 + ChatGPT + Zero-Shot — 82.3 (2023-07-14) C3 + ChatGPT + Zero-Shot — 82.3 (2023-07-14) DAIL-SQL + GPT-4 + Self-Consistency — 86.6 (2023-08-29) DAIL-SQL + GPT-4 + Self-Consistency — 86.6 (2023-08-29) PET-SQL — 87.6 (2024-03-13) PET-SQL — 87.6 (2024-03-13) MSc-SQL — 84.7 (2024-10-16) MSc-SQL — 84.7 (2024-10-16) MARLO + Claude 2.1 — 84.0 (2024-10-17) MARLO + Claude 2.1 — 84.0 (2024-10-17) XiYan-SQL — 89.65 (2024-11-13) XiYan-SQL — 89.65 (2024-11-13) X-Linking — 84.9 (2025-09-07) LitE-SQL — 88.45 (2025-10-10) AGENTIQL — 86.07 (2025-10-12) MARS-SQL — 89.75 (2025-11-02) AutoLink — 91.2 (2025-11-21) DSR-SQL — 68.32 (2025-11-26) Single-Agent — 85.5 (2026-01-25) AV-SQL — 85.59 (2026-04-08) SIRIUS-SQL — 91.2 (2026-05-31) ZAS-SQL — 87.2 (2026-06-06) CoTE-SQL — 79.6 (2026-06-14) DecoSearch — 88.31 (2026-06-16) Policy-Conditioned Constrained Decoding — 88.7 (2026-07-14) AttnLink — 95.95 (2026-08-01) SERL-SQL — 89.92 (2026-08-01) MERIT — 69.79 (2026-08-06) T5-3B + PICARD — 75.1 (2021-09-10) Graphix-3B+PICARD — 77.6 (2023-01-18) RESDSQL-3B + NatSQL — 79.9 (2023-02-12) DIN-SQL + GPT-4 — 85.3 (2023-04-21) DAIL-SQL + GPT-4 + Self-Consistency — 86.6 (2023-08-29) PET-SQL — 87.6 (2024-03-13) XiYan-SQL — 89.65 (2024-11-13) MARS-SQL — 89.75 (2025-11-02) AutoLink — 91.2 (2025-11-21) AttnLink — 95.95 (2026-08-01)
RankModel Execution Accuracy (Test)Exact Match Accuracy (Test)Execution Accuracy (Dev)Exact Match Accuracy (Dev) PaperCodeYear
1 AttnLink 자동 추출 95.95––– AttnLink: Turning Attention into Schema Links for Text-to-SQL 2026
2 SIRIUS-SQL 자동 추출 91.20––– SIRIUS-SQL: Anchoring Multi-Candidate Text-to-SQL in Execution Feedback 2026
2 AutoLink 자동 추출 91.2––– AutoLink: Autonomous Schema Exploration and Expansion for Scalable Schema Linking in Text-to-SQL at Scale 2025
4 SERL-SQL 자동 추출 89.92––– SERL-SQL: Selective Hindsight Distillation for Text-to-SQL Reinforcement Agentic Learning 2026
5 MARS-SQL 자동 추출 89.75––– MARS-SQL: A multi-agent reinforcement learning framework for Text-to-SQL 2025
6 XiYan-SQL 89.65––– A Preview of XiYan-SQL: A Multi-Generator Ensemble Framework for Text-to-SQL XGenerationLab/XiYan-SQL · xgenerationlab/xiyan_mcp_server · XGenerationLab/M-Schema · +2 2024
7 Policy-Conditioned Constrained Decoding 자동 추출 88.7––– Policy-Conditioned Constrained Decoding for Column-Level Access Control in Text-to-SQL arXivSub_daily_arxiv 2026
8 LitE-SQL 자동 추출 88.45––– LitE-SQL: A Lightweight and Efficient Text-to-SQL Framework with Vector-based Schema Linking and Execution-Guided Self-Correction 2025
9 DecoSearch 자동 추출 88.31––– DecoSearch: Complexity-Aware Routing and Plan-Level Repair for Text-to-SQL 2026
10 PET-SQL 87.666.6–– PET-SQL: A Prompt-Enhanced Two-Round Refinement of Text-to-SQL with Cross-consistency zhshlii/petsql 2024
11 ZAS-SQL 자동 추출 87.2––– ZAS-SQL: Distilling Rules from Failures for Zero-Shot Text-to-SQL 2026
12 DAIL-SQL + GPT-4 + Self-Consistency 86.6–84.474.4 Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation beachwang/dail-sql 2023
13 AGENTIQL 자동 추출 86.07––– AGENTIQL: An Agent-Inspired Multi-Expert Framework for Text-to-SQL Generation 2025
14 AV-SQL 자동 추출 85.59––– AV-SQL: Decomposing Complex Text-to-SQL Queries with Agentic Views 2026
15 Single-Agent 자동 추출 85.5––– LLM-Based SQL Generation: Prompting, Self-Refinement, and Adaptive Weighted Majority Voting 2026
16 DIN-SQL + GPT-4 85.360–– DIN-SQL: Decomposed In-Context Learning of Text-to-SQL with Self-Correction mohammadrezapourreza/few-shot-nl2sql-with-prompting 2023
17 X-Linking 자동 추출 84.9––– X-SQL: Expert Schema Linking and Understanding of Text-to-SQL with Multi-LLMs 2025
18 MSc-SQL 84.7––– MSc-SQL: Multi-Sample Critiquing Small Language Models For Text-To-SQL Translation layer6ai-labs/msc-sql 2024
19 MARLO + Claude 2.1 84.0–83.6– Learning Metadata-Agnostic Representations for Text-to-SQL In-Context Example Selection 2024
20 C3 + ChatGPT + Zero-Shot 82.3–81.8– C3: Zero-shot Text-to-SQL with ChatGPT bigbigwatermalon/c3sql 2023
1–20 / 56 다음 → 페이지당 10 20 50 100