News Hub

GLiDE family includes decision models ranging from 0.4 billion to 27 billion parameters.

Panel discussion shares lessons from deploying open-weight models in production environments.

Guide to agent memory solutions for building agents that retain information.

Fastino introduces a training API for supervised fine-tuning and reinforcement learning.

Article compares three methods for adapting open weight models: fine-tuning, retrieval-augmented generation, and prompt engineering.

Fastino introduced GLiDE, described as the first thinking decision model.

GLiNER2.5-Decide is an open-weight model designed for structured decision-making tasks.

Article discusses the best open weight models released during summer 2026.

GLiNER2.5 performs efficient span-free information extraction using schema-driven interface.

Guide on using open-weight AI models for developers in 2026.

Fastino released finance and healthcare specialized versions of Nemotron 3.5 Lightning.

Fastino published guidance on fine-tuning open-weight models covering selection, data curation, strategies, and evaluation.

Fastino argues loop engineering requires a smarter inference layer.

Fastino examines performance gap between open-weight and proprietary frontier language models.

Hermes Agent is a self-improving AI agent framework.

GLiNER2-Guardrails-PII-Multi is a small language model for safety moderation and privacy filtering.

OpenCode is a complete guide to open source AI coding agents as of 2026.

A guide to LLM inference techniques and methods.

Guide to choosing the best coding models in 2026.

Guide to small language models.