Feed · Updated 2026-05-07
Latest
Updated 2026-05-07 — what's new in local AI on this site. Workflows and system pages reviewed in the last 60 days, tools and hardware freshly enriched, newest benchmarks.
Looking for sprint-level history? See the changelog or subscribe via RSS.
Recently reviewed
18Workflows and system pages whose lastReviewed is within 60 days. Hand-revised, not auto-touched.
- WorkflowHomelab AI API gateway
- System guideLocal AI on Linux — operator-grade deployment paths
- WorkflowLocal coding-agent system
- WorkflowLocal evaluation lab
- WorkflowLocal fine-tuning workstation
- WorkflowLocal voice assistant pipeline
- System guideMaintaining a local-AI build over time
- WorkflowMulti-user local AI server
- System guideObservability for local AI
- WorkflowOffline RAG pipeline
- WorkflowPrivate ChatGPT replacement
- WorkflowPrivate job-search assistant
- System guideQuantization formats for local AI — GGUF, AWQ, GPTQ, EXL2, FP8, MLX
- System guideSecuring a local-AI deployment
- System guideHow agent execution systems actually work
- System guideWhat agent memory actually is
- System guideWhat distributed inference actually is
- System guideWhat MCP is really solving
Freshly enriched
29Tool and hardware pages with new long-form L1.25 review bodies in the last 60 days.
- ToolMLX-LM
- ToolExLlamaV2
- Toolllama.cpp
- ToolOllama
- ToolSGLang
- ToolOpen WebUI
- ToolLibreChat
- ToolIntel OpenVINO
- ToolLM Studio
- ToolTensorRT-LLM
- ToolvLLM
- ToolONNX Runtime
- ToolROCm
- ToolAnythingLLM
- HardwareNVIDIA GeForce RTX 3090
- HardwareNVIDIA GeForce RTX 4070 Ti Super
- HardwareApple M3 Ultra
- HardwareQualcomm Snapdragon X Elite
- HardwareAMD Radeon RX 7900 XTX
- HardwareApple A18 Pro
- HardwareAMD Ryzen AI 9 HX 370 (Strix Point)
- HardwareIntel Core Ultra 7 258V (Lunar Lake)
- HardwareNVIDIA GeForce RTX 5080
- HardwareNVIDIA GeForce RTX 5090
- HardwareIntel Arc A770 16GB
- HardwareApple M4 Max
- HardwareNVIDIA GeForce RTX 3060 12GB
- HardwareNVIDIA GeForce RTX 4090
- HardwareNVIDIA H100 SXM
Newest benchmarks
10The 10 most recent benchmark rows by run date. See the full table at /benchmarks.
- Benchmark · communityLlama 3.1 8B Instruct on Apple M3 Max
- Benchmark · communityLlama 3.3 70B Instruct on Apple M3 Ultra
- Benchmark · communityLlama 3.3 70B Instruct on NVIDIA GeForce RTX 4090
- Benchmark · communityLlama 3.1 8B Instruct on NVIDIA GeForce RTX 4090
- Benchmark · communityLlama 3.1 8B Instruct on NVIDIA GeForce RTX 3090
- Benchmark · communityLlama 3.1 8B Instruct on NVIDIA GeForce RTX 5090
- Benchmark · ownerLlama 3.1 8B Instruct on NVIDIA GeForce RTX 5080
- Benchmark · ownerQwen 2.5 Coder 7B Instruct on NVIDIA GeForce RTX 3080 16GB (Mobile)
- Benchmark · communityQwen 2.5 Coder 32B Instruct on NVIDIA GeForce RTX 4090
- Benchmark · communityDeepSeek R1 Distill Qwen 32B on NVIDIA GeForce RTX 4090