LLMs

723 items

RESEARCHarXiv CS.CL·4/20/2026

"Excuse me, may I say something..." CoLabScience, A Proactive AI Assistant for Biomedical Discovery and LLM-Expert Collaborations

CoLabScience is introduced as a proactive LLM assistant aimed at accelerating biomedical discovery by facilitating collaborations between AI and human experts. It features PULI, a novel reinforcement learning framework for timely interventions in scientific discussions, and also presents BSDD, a new benchmark dataset of simulated research dialogue.

LLMs AI collaboration reinforcement learning datasets

ARTICLEDEV.to AI·4/26/2026

I Watched Google Cloud NEXT '26 ~ Here Is What Actually Matters for Developers

An AI agent engineer, Aniruddha Adak, provides an honest take on Google Cloud NEXT 2026, focusing on what genuinely matters for developers regarding agentic systems and LLMs, filtering out marketing hype. He shares his personal insights after thoroughly watching the keynotes.

LLMs cloud computing AI developer tools

ARTICLEDEV.to AI·24d ago

Inside Naver Cloud: The AI Access Pioneer the West Hasn't Noticed

While global tech giants fiercely compete in the AI 'Gold Rush', South Korea's Naver Cloud is quietly building a unique, full-stack AI ecosystem. This initiative aims to democratize access to powerful LLMs and AI services, focusing on efficiency and localization rather than just scale.

AI Democratization LLMs AI Naver Cloud

ARTICLEDEV.to AI·5/4/2026

The Hidden Cost of MCP: Why AI Agents Need Code Mode

The content discusses the hidden cost of AI agents carrying entire tool catalogs in context, leading to increased costs, latency, and errors. It proposes "Code Mode" as a solution, enabling agents to discover tools as needed and orchestrate workflows via code, thereby reducing context and optimizing costs in production.

AI architecture LLMs code mode cost optimization

RESEARCHDEV.to AI·5/7/2026

GPT-4.1 Hits 24.65% Derm Accuracy on Real Cases vs 42.25% Benchmarks

A new study reveals that multimodal large language models (LLMs) like GPT-4.1 exhibit a significant drop in diagnostic accuracy when applied to real hospital dermatology cases, compared to public benchmarks. The research, encompassing 5,811 cases, found GPT-4.1 achieved 24.65% accuracy in real clinical settings versus 42.25% on benchmarks.

LLMs Healthcare Dermatology benchmarking

ARTICLEDEV.to AI·5/1/2026

Building Production-Grade Tools for AI Agents: What Works After 100 Deployments

This article argues that tool design is more critical than prompt engineering for the reliability of production AI agents, preventing hallucinations at a structural level. Drawing from over 100 deployments, it promises to share concrete patterns, code examples, and anti-patterns for building robust AI tools.

LLMs Production AI tool design AI agents

DOCDEV.to AI·17d ago

Building a cost-efficient LLM caching layer in Python

This tutorial details building a cost-efficient LLM caching layer in Python to reduce API costs. It uses exact-match via Redis and semantic near-duplicate detection via cosine similarity. This approach can lead to significant monthly savings by preventing redundant API calls.

LLMs Redis cost optimization Caching

ARTICLEDEV.to AI·5/1/2026

Your AI Agent Forgets Everything Between Sessions (Here's How to Fix It)

This article discusses the problem of AI agents forgetting progress between sessions, leading to wasted time re-explaining context. Reloading conversation history is ineffective due to its massive size, excessive noise, and lack of composability.

LLMs memory management AI agents

RESEARCHDEV.to AI·5/7/2026

ReTool: Reinforcement Learning for Strategic Tool Use in LLMs

ReTool introduces a novel reinforcement learning framework designed to enhance the strategic tool-use capabilities of Large Language Models. This approach aims to improve how LLMs select and utilize external tools to solve complex tasks more effectively and efficiently.

LLMs reinforcement learning machine learning tool use

RESEARCHarXiv CS.CL·4/21/2026

QU-NLP at QIAS 2026: Multi-Stage QLoRA Fine-Tuning for Arabic Islamic Inheritance Reasoning

This paper presents QU-NLP's multi-stage QLoRA fine-tuning strategy for Arabic Islamic inheritance reasoning using Qwen3-4B. The model achieved a 90% MIR-E score, demonstrating competitive performance with minimal computational resources.

LLMs Legal AI Arabic AI NLP

ARTICLEDEV.to AI·5/10/2026

How To Select an Enterprise LLM

The article discusses the intensifying competition in enterprise LLM deployment, highlighting new models from OpenAI and Mistral AI. It emphasizes the need for a systematic benchmarking approach that considers latency, cost, and task-specific performance, urging organizations to use a multi-phase evaluation framework to align models with business objectives.

LLMs model selection benchmarking AI deployment

ARTICLEDEV.to AI·4/30/2026

Building an AI Agent Harness from Scratch: The Architecture Between LLM and Agent

The article emphasizes the crucial role of the 'harness' or orchestration layer in transforming an LLM into a reliable AI agent. It argues that this layer, which manages tools and state, is what truly differentiates a robust agent from a simple conversational AI.

LLMs orchestration Architecture AI agents

DOCDEV.to AI·4/21/2026

Building a C# Agent with Microsoft Agent Framework and Ollama

This content details building a C# agent with Microsoft Agent Framework 1.0 and Ollama, deployed on Azure Container Apps. It highlights the framework's significantly simpler API compared to its predecessors like Semantic Kernel, streamlining agent development.

LLMs Microsoft Agent Framework Ollama C++

DOCDEV.to AI·5/4/2026

How to Use MCP Servers With Ollama and Local LLMs

This guide explains how to use MCP servers with local LLM models via Ollama, which does not natively implement the MCP protocol. It requires setting up a bridge to communicate between the Ollama API and the MCP protocol for tool call dispatch.

AI integration MCP Protocol LLMs Ollama

DOCfreeCodeCamp (YouTube)·5/7/2026

Open Models Coding Essentials – Running LLMs Locally and in the Cloud Course

This course covers the coding essentials for deploying and running Large Language Models (LLMs) locally and in cloud environments. It provides practical knowledge on managing these models in various operational setups.

LLMs cloud computing learning open models

Open Models Coding Essentials – Running LLMs Locally and in the Cloud Course

ARTICLEDEV.to AI·25d ago

The Livingrimoire advantage: a tiny “welcome back” skill that LLMs can’t match

The article introduces the "DiOkaeri" skill from Livingrimoire, demonstrating how it provides specific, time-dependent behavioral responses that large language models (LLMs) cannot reliably replicate. It emphasizes that robust AI agent development requires dedicated architecture for behavior generation, moving beyond text-only LLM capabilities.

AI architecture LLMs Livingrimoire behavioral AI

ARTICLEDEV.to AI·26d ago

The primary reader changed

This article explains how AI agents read code with a fundamentally different approach compared to humans, leading to a 7.5x increase in token cost for the same features. This shift in the "primary reader" introduces new cost dynamics and necessitates a re-evaluation of existing programming patterns.

LLMs Software development AI programming

ARTICLEDEV.to AI·28d ago

Stop feeding raw HTML to your LLMs (Solving the Agentic Token Tax)

Autonomous AI agents interacting with the web are burdened by an 'Agentic Token Tax' due to inefficiently feeding raw HTML to LLMs, leading to high API costs and latency. This approach often fails with modern web applications; thus, a deterministic protocol like Web Speed is proposed as a more robust solution than better scrapers.

LLMs Web Interaction token optimization web-scraping

DOCDEV.to AI·4/28/2026

Chinese AI APIs Are 80% Cheaper Than OpenAI — I Tested Them All and Built an Integration Guide

The author tested Chinese AI APIs like GLM-5.1, Qwen, and DeepSeek, finding them 80% cheaper than OpenAI with comparable quality and strong multilingual support. He offers a comprehensive integration guide with code, pricing comparisons, and setup instructions for easy migration.

Chinese AI models integration LLMs AI APIs

ARTICLEDEV.to AI·26d ago

Giving AI agents knowledge they were never trained on

The docs-mcpserver tool enables AI agents to access specific knowledge, such as internal specifications or particular library versions, by reading various documentation formats like Markdown, API references, and schemas. This helps LLMs code accurately without guessing APIs or wasting tokens on extensive parsing.

LLMs Software development documentation knowledge base