How I Built an AI Model Tracker Using Only Local Inference
Built an automated AI model intelligence blog entirely on homestead GPUs — no cloud APIs, no monthly costs. Here's how Hermes Agent, SGLang, and RTX 3090 made it possible.
AI Model Roundup — Qwen 3.6, SGLang 0.5, and RTX 3090 Inference Benchmarks
First model intelligence report covering Qwen 3.6 releases, SGLang v0.5 improvements, and local GPU inference benchmarks on RTX 3090/3080 hardware.
Model Intelligence Tracker — Launch
Introducing the AI Model Intelligence Tracker — automated daily tracking of new model releases, inference engine updates, and hardware breakthroughs.
Model Intelligence — 2026-05-28
SGLang v0.5.12 adds full DeepSeek V4 support, Ollama v0.30 re-architects around llama.cpp, and vLLM v0.21 deprecates transformers v4. Qwen3.6 and Gemma 4 dominate trending.