Running Local LLMs on AMD and Intel GPUs in 2026
AMD's ROCm support splits by card and by OS, Intel archived both of its own LLM libraries this year, and llama.cpp is the path still standing.
Tag
#local-ai
Serving a Local Model to Your Household Without Exposing It
None of the mainstream inference servers ask for a password, and vLLM binds every interface by default. The safe pattern for household serving.
What Quantization Actually Costs You in Quality
8-bit measures close to lossless. 4-bit ranges from a 0.6% gain to a 59% drop on one benchmark, depending on the model. The figures, fully attributed.
Which Open Models You Can Actually Use Commercially
Open weights is not an open licence. Verified Hugging Face licence tags for 46 model repos, and the size and version traps that block shipping.
Qwen3.8-27B Ships Apache 2.0; AMD Adds Day-0 Support
Qwen's 27B vision-language model is on Hugging Face ungated under Apache 2.0, and AMD says it needs roughly 24GB of VRAM to run comfortably.
Local AI Without a GPU: What Actually Runs (August 2026)
Which open models run on CPU and RAM alone, how to size them, why mixture-of-experts helps, and what a machine with no discrete GPU cannot do.
Leading Frontier AI Models: The Capability Ceiling (August 2026)
What sits at the top of the field right now, hosted or downloadable, with every licence, price and index score re-read from a primary source today.
Best GPUs for Local AI (August 2026)
VRAM decides what runs at all, bandwidth decides how fast it types. A buying guide by tier across NVIDIA, AMD, Intel, Apple and the used market.
Best Local LLM Runners (August 2026)
Ollama, LM Studio, Jan, llama.cpp and vLLM compared: licences, engines, OpenAI-compatible APIs, multi-user serving, and which to install first.
Best Local RAG Setups (August 2026)
Six self-hosted RAG stacks compared: AnythingLLM, Open WebUI, LibreChat, Msty, RAGFlow, Cherry Studio. Licences, embedders, and who can rerank.
Best Open-Weight Models, Ranked by Capability (August 2026)
Sixteen downloadable-weight models ranked by capability and licence, with every parameter count and licence re-checked against the Hugging Face API.
Liquid AI Lands on Setapp While iOS Private Relay Leaks Real IPs
Liquid AI's 2.6B model ships via Setapp for offline Mac agents, while iCloud Private Relay leaks real IPs through passkeys in the same 24 hours.
Best Local Embedding and Reranker Models by VRAM (September 2026)
Local embedding and reranking models for RAG, sized by VRAM. Real GGUF file sizes, licences, and the runtime gap that stops Ollama reranking.
Best Local Image Generation Models by VRAM (2026)
Open image models sized for real GPUs, with the text encoder counted in and the licence checked. FLUX, Z-Image, Qwen-Image and SD3.5 compared.
ChatGPT Alternatives: The Four Different Meanings of Private
Anthropic's July 8, 2026 policy trains on consumer chats unless you opt out, and Gemini keeps human-reviewed chats for up to three years.
6GB VRAM: What You Can Actually Run Locally (2026)
Local AI on a 6GB GPU: GTX 1660, RTX 2060, RTX 3050 and laptop cards. Real weight sizes for chat, coding, vision, speech, translation and RAG.
Local AI by VRAM: Which Models Fit Your GPU (September 2026)
Find your VRAM tier, then the current open-weight pick for chat, coding, vision, speech, translation or agents. Eighteen guides, one index.
Run an LLM Locally on Android and iPhone (September 2026)
Apple caps its on-device model at 4096 tokens per session, and Google's own tables put Gemma 4 E2B at 25.0 decode tokens/sec on an iPhone 17 Pro CPU.
Run LLMs Locally on a Mac: What Actually Fits (September 2026)
Apple Silicon has no discrete VRAM, so tier guides mislead Mac owners. The real ceilings are bandwidth and the GPU-usable slice of unified memory.
GGUF vs AWQ vs GPTQ vs MLX: Which Quantization to Use
Quantization formats compared with real file sizes and bits per weight. Why Q4 does not halve a model, and why low quants generate faster.
Qwen3.8-Max Goes Open Weight; 27B Targets 17GB
Alibaba announced 2.4T-parameter Qwen3.8-Max open weights and a 27B sibling that Unsloth says will run in 17GB of RAM or VRAM.
MiniMax-M3 Lands in llama.cpp: Sparse Attention and Vision
Two back-to-back merges add MiniMax Sparse Attention and a Qwen2.5-VL style vision tower to llama.cpp, but every existing MiniMax-M3 GGUF must be regenerated.
Open-Weight AI Coalition Draws a Line on Distillation
A new industry letter asks Washington to protect open-weight AI while separating legitimate distillation from alleged theft of closed models.
Cisco's Open Antares Models Let You Audit Code Locally
Cisco released Antares-350M and Antares-1B open-weight models for vulnerability localization. They run locally and cost about 172x less than GPT-5.5.