arXiv:2512.08193v1 Announce Type: new 
Abstract: We present ClinicalTrialsHub, an interactive search-focused platform that consolidates all data from ClinicalTrials.gov and augments it by automatically extracting and structuring trial-relevant information from PubMed research articles. Our system effectively increases access to structured clinical trial data by 83.8% compared to relying on ClinicalTrials.gov alone, with potential to make access easier for patients, clinicians, researchers, and policymakers, advancing evidence-based medicine. ClinicalTrialsHub uses large language models such as GPT-5.1 and Gemini-3-Pro to enhance accessibility. The platform automatically parses full-text research articles to extract structured trial information, translates user queries into structured database searches, and provides an attributed question-answering system that generates evidence-grounded answers linked to specific source sentences. We demonstrate its utility through a user study involving clinicians, clinical researchers, and PhD students of pharmaceutical sciences and nursing, and a systematic automatic evaluation of its information extraction and question answering capabilities.

أطلق ClinicalTrialsHub منصة تفاعلية تدمج بيانات ClinicalTrials.gov وتستخرج معلومات ذات صلة من مقالات PubMed، مما يعزز الوصول إلى بيانات التجارب السريرية بنسبة 83.8%. تستخدم هذه الأداة المبتكرة نماذج لغوية متقدمة لتسهيل عمليات البحث المنظمة وتقديم إجابات قائمة على الأدلة لاستفسارات المستخدمين.

ClinicalTrialsHub ha lanzado una plataforma interactiva que integra datos de ClinicalTrials.gov y extrae información relevante de artículos de PubMed, mejorando el acceso a datos de ensayos clínicos en un 83.8%. Esta herramienta innovadora utiliza modelos de lenguaje avanzados para facilitar búsquedas estructuradas y proporcionar respuestas basadas en evidencia a las consultas de los usuarios.

ClinicalTrialsHub a lancé une plateforme interactive qui intègre les données de ClinicalTrials.gov et extrait des informations pertinentes d'articles de PubMed, améliorant l'accès aux données des essais cliniques de 83,8 %. Cet outil innovant utilise des modèles linguistiques avancés pour faciliter les recherches structurées et fournir des réponses basées sur des preuves aux requêtes des utilisateurs.

ClinicalTrialsHub has launched an interactive platform that integrates data from ClinicalTrials.gov and extracts relevant information from PubMed articles, enhancing access to clinical trial data by 83.8%. This innovative tool utilizes advanced language models to facilitate structured searches and provide evidence-based answers to user queries.

ClinicalTrialsHub: Bridging Registries and Literature for Comprehensive Clinical Trial Access

<img width="800" height="450" src="https://analyticsindiamag.com/wp-content/uploads/2023/04/databricks-1300x731.jpg" class="webfeedsFeaturedVisual wp-post-image" alt="" style="display: block; margin: auto; margin-bottom: 5px;max-width: 100%;" link_thumbnail="" decoding="async" loading="lazy" srcset="https://analyticsindiamag.com/wp-content/uploads/2023/04/databricks-1300x731.jpg 1300w, https://analyticsindiamag.com/wp-content/uploads/2023/04/databricks-768x432.jpg 768w, https://analyticsindiamag.com/wp-content/uploads/2023/04/databricks-1536x864.jpg 1536w, https://analyticsindiamag.com/wp-content/uploads/2023/04/databricks.jpg 1920w" sizes="auto, (max-width: 800px) 100vw, 800px" />On the benchmark, Anthropic’s Claude Opus 4.5 Agent solved 37.4% whereas OpenAI’s GPT-5.1 Agent scored 43.1% on the full data set.
The post <a href="https://analyticsindiamag.com/ai-news-updates/databricks-benchmark-tests-ai-on-enterprise-tasks-that-demand-unforgiving-accuracy/">Databricks Benchmark Tests AI on Enterprise Tasks That Demand ‘Unforgiving Accuracy’</a> appeared first on <a href="https://analyticsindiamag.com">Analytics India Magazine</a>.

أجرت شركة Databricks اختبارات مرجعية على نماذج الذكاء الاصطناعي، كاشفةً أن وكيل Claude Opus 4.5 من Anthropic حقق درجة 37.4%، بينما حصل وكيل GPT-5.1 من OpenAI على 43.1% في مهام الشركات التي تتطلب دقة عالية. تسلط هذه التقييمات الضوء على المشهد التنافسي في أداء الذكاء الاصطناعي، خاصة في التطبيقات التجارية.

Databricks realizó pruebas de referencia en modelos de IA, revelando que el agente Claude Opus 4.5 de Anthropic obtuvo un puntaje del 37.4%, mientras que el agente GPT-5.1 de OpenAI alcanzó un 43.1% en tareas empresariales que requieren alta precisión. Esta evaluación resalta el paisaje competitivo en el rendimiento de la IA, especialmente en aplicaciones empresariales.

Databricks a réalisé des tests de référence sur des modèles d'IA, révélant que l'agent Claude Opus 4.5 d'Anthropic a obtenu un score de 37,4 %, tandis que l'agent GPT-5.1 d'OpenAI a atteint 43,1 % sur des tâches d'entreprise nécessitant une grande précision. Cette évaluation met en lumière le paysage concurrentiel en matière de performance de l'IA, en particulier dans les applications d'entreprise.

Databricks conducted benchmark tests on AI models, revealing that Anthropic’s Claude Opus 4.5 Agent achieved a score of 37.4%, while OpenAI’s GPT-5.1 Agent scored 43.1% on enterprise tasks requiring high accuracy. This assessment highlights the competitive landscape in AI performance, particularly in enterprise applications.

Databricks Benchmark Tests AI on Enterprise Tasks That Demand ‘Unforgiving Accuracy’

<img src="https://res.infoq.com/news/2025/12/openai-gpt-51/en/headerimage/generatedHeaderImage-1765114067009.jpg"/>OpenAI recently released upgrades to their GPT-5 model. GPT‑5.1 Instant, the default chat model, has improvements to instruction following. GPT‑5.1 Thinking, the reasoning model, is faster and gives more understandable responses. GPT‑5.1-Codex-Max, the coding model, is trained to use compaction to perform long-running tasks. By Anthony Alford

أطلقت OpenAI تحديثات لنموذج GPT-5 الخاص بها، حيث قدمت GPT-5.1 Instant لتحسين اتباع التعليمات، وGPT-5.1 Thinking للتفكير الأسرع، وGPT-5.1-Codex-Max لتحسين قدرات البرمجة. تهدف هذه التحديثات إلى تعزيز تفاعل المستخدم وجودة الاستجابة في تطبيقات الذكاء الاصطناعي.

OpenAI ha lanzado actualizaciones a su modelo GPT-5, introduciendo GPT-5.1 Instant para una mejor comprensión de instrucciones, GPT-5.1 Thinking para un razonamiento más rápido y GPT-5.1-Codex-Max para capacidades de codificación mejoradas. Estas actualizaciones buscan mejorar la interacción del usuario y la calidad de las respuestas en aplicaciones de IA.

OpenAI a lancé des mises à jour de son modèle GPT-5, introduisant GPT-5.1 Instant pour une meilleure compréhension des instructions, GPT-5.1 Thinking pour un raisonnement plus rapide, et GPT-5.1-Codex-Max pour des capacités de codage améliorées. Ces mises à jour visent à améliorer l'interaction utilisateur et la qualité des réponses dans les applications d'IA.

OpenAI has launched upgrades to its GPT-5 model, introducing GPT-5.1 Instant for improved instruction following, GPT-5.1 Thinking for faster reasoning, and GPT-5.1-Codex-Max for enhanced coding capabilities. These updates aim to enhance user interaction and response quality in AI applications.

OpenAI's New GPT-5.1 Models are Faster and More Conversational

arXiv:2509.24282v2 Announce Type: replace 
Abstract: Large Language Model (LLM) agents excel at multi-step, tool-augmented tasks. However, smart homes introduce distinct challenges, requiring agents to handle latent user intents, temporal dependencies, device constraints, scheduling, and more. The main bottlenecks for developing smart home agents with such capabilities include the lack of a realistic simulation environment where agents can interact with devices and observe the results, as well as a challenging benchmark to evaluate them. To address this, we introduce $\textbf{SimuHome}$, a time-accelerated home environment that simulates smart devices, supports API calls, and reflects changes in environmental variables. By building the simulator on the Matter protocol, the global industry standard for smart home communication, SimuHome provides a high-fidelity environment, and agents validated in SimuHome can be deployed on real Matter-compliant devices with minimal adaptation. We provide a challenging benchmark of 600 episodes across twelve user query types that require the aforementioned capabilities. Our evaluation of 16 agents under a unified ReAct framework reveals distinct capabilities and limitations across models. Models under 7B parameters exhibited negligible performance across all query types. Even GPT-4.1, the best-performing standard model, struggled with implicit intent inference, state verification, and particularly temporal scheduling. While reasoning models such as GPT-5.1 consistently outperformed standard models on every query type, they required over three times the average inference time, which can be prohibitive for real-time smart home applications. This highlights a critical trade-off between task performance and real-world practicality.

تم تقديم SimuHome كمعيار مصمم لتقييم وكلاء نماذج اللغة الكبيرة (LLM) في المنازل الذكية، حيث يتناول تحديات مثل نية المستخدم، والاعتماد الزمني، وقيود الأجهزة. يوفر هذا البيئة المعجلة زمنياً محاكاة للأجهزة الذكية ويدعم استدعاءات API، مما يوفر منصة واقعية لتفاعل الوكلاء.

SimuHome se ha presentado como un benchmark diseñado para evaluar agentes de modelos de lenguaje (LLM) en hogares inteligentes, abordando desafíos como la intención del usuario, las dependencias temporales y las limitaciones de los dispositivos. Este entorno acelerado en el tiempo simula dispositivos inteligentes y admite llamadas API, proporcionando una plataforma realista para la interacción de los agentes.

SimuHome a été introduit comme un benchmark conçu pour évaluer les agents de modèles de langage (LLM) dans les maisons intelligentes, en abordant des défis tels que l'intention des utilisateurs, les dépendances temporelles et les contraintes des appareils. Cet environnement accéléré dans le temps simule des appareils intelligents et prend en charge les appels API, offrant une plateforme réaliste pour l'interaction des agents.

SimuHome has been introduced as a benchmark designed for evaluating smart home large language model (LLM) agents, addressing challenges such as user intent, temporal dependencies, and device constraints. This time-accelerated environment simulates smart devices and supports API calls, providing a realistic platform for agent interaction.

ClinicalTrialsHub: Bridging Registries and Literature for Comprehensive Clinical Trial Access

Was this article worth reading? Share it