Organizations that cannot send data to a cloud API increasingly ask: how good is Text-to-SQL if the model must run on-premises on open weights, and which popular accuracy "recipes" are worth their compute? We answer with an honest, fully reproducible benchmark on the BIRD development split (n=1534, Execution Accuracy), evaluating three open model families across two generations -- Qwen2.
arXiv:2606. 29733v1 Announce Type: cross Abstract: Organizations that cannot send data to a cloud API increasingly ask: how good is Text-to-SQL if the model must run on-premises on open weights, and which popular accuracy "recipes" are worth their compute?
arXiv:2607. 06799v1 Announce Type: cross Abstract: Evaluating uncertainty in AI-generated SQL queries requires estimating whether a query is correct, where correct means it executes to the same result as a human-written reference.
ESQ‑Bench is an Oracle‑first NL2SQL benchmark that introduces systematic complexity tiers and silent‑divergence evaluation across three enterprise schema levels. It provides six populated schemas (465 tables, 164,682 rows) on Oracle, PostgreSQL, MySQL, and SQL Server, along with 550 gold‑validated question‑query pairs and a four‑metric evaluation harness. The benchmark reveals that state‑of‑the‑art models such as GPT‑4o and Claude Sonnet 4.6 experience significant drops in execution accuracy and high silent‑divergence rates as schema complexity increases, highlighting a gap between closed‑API models and open‑weight baselines on enterprise Oracle schemas.
SemVerBench is a benchmark that evaluates how well large language models (LLMs) understand and apply version-constraint resolution semantics, such as determining whether a version satisfies constraints like ^1.2.3 or >=2.0. The study finds that many models struggle with certain corner cases, with GPT‑5.1 performing poorly while Claude and Opus perform much better. The authors suggest that the failures stem from an activation/application gap rather than a lack of knowledge, and recommend that coding agents delegate version resolution to a dedicated resolver tool.
The paper evaluates nine on‑device named‑entity recognition models ranging from classical taggers to large language models, measuring not only accuracy but also latency and output validity. Using a silver‑gold benchmark derived from an LLM judge panel and a human‑validated corpus, the study shows that encoder‑based models achieve comparable accuracy to a 4 B instruct LLM while being much smaller, faster, and producing no malformed output. Confidence calibration of GLiNER is analyzed, revealing over‑confidence but improved reliability after temperature scaling and thresholding.
The paper introduces LAST-CQ, a five-agent, training‑free, execution‑grounded framework for Text‑to‑Cypher that evaluates which components of an agentic pipeline contribute most to performance. Experiments on 2,471 live‑database queries across six backbones show that removing correction reduces execution‑BLEU by 3.1–12.3%, while substituting schema‑grounded feedback with raw error strings has negligible impact. Parallel sampling degrades quality by 10–11%, whereas failure detection and retry routing recover 91.7% of initially failed queries, highlighting that simple failure handling is more effective than sophisticated feedback or increased sampling.
The paper introduces DRL, a deterministic relational middleware layer designed to enable transaction-safe natural‑language to SQL (NL2SQL) interfaces over large enterprise OLTP catalogs. DRL interposes between front‑ends and SQL back‑ends, employing dynamic context pruning, relational AST typing, and transactional safeguards (EXPLAIN gating and NULL guards) to keep context within LLM attention limits and detect silent divergence. Experiments on PostgreSQL and MySQL show significant context reductions (up to 92%) and high execution match rates (≈53%) for GPT‑4o, Claude Sonnet 4.5, and Gemini 2.5 Flash, while also revealing that evaluation code quality can materially affect reported performance gaps.
arXiv:2609.05637v2 Announce Type: replace Abstract: A popular way to improve Retrieval-Augmented Generation (RAG) is to rewrite the user's question into several variants and search with all of them....
arXiv:2607. 25583v1 Announce Type: new Abstract: Parameter-efficient fine-tuning (PEFT) and low-bit quantization are now standard tools for adapting language models under tight compute budgets, yet their interaction is most often studied on billion-parameter models where the design space is expensive to explore.
arXiv:2608. 11889v1 Announce Type: cross Abstract: Prompting-based (\textit{i}.
The paper evaluates a manager‑worker scaffold that uses a shared filesystem workspace to orchestrate multi‑agent large language model (LLM) coding tasks without training or tuning. Across nine models—including five open‑weight and four closed‑weight systems—the scaffold yields statistically significant accuracy gains for some models (e.g., Qwen3.8‑27B, GPT‑5.6‑Luna, GPT‑5.6‑Terra, Kimi‑K3, Minimax‑M3) while producing null or negative effects for others (e.g., Qwen3.6‑35B). The study shows that the manager can triple token usage but still achieves higher accuracy at a fraction of the cost compared to larger single‑pass models, with key mechanisms identified as context management and problem decomposition.
arXiv:2606. 08481v1 Announce Type: cross Abstract: Enterprise property graphs vary widely in schema structure, internal terminology, domain assumptions, governance constraints, and user interaction patterns.