Search

Few direct matches — filled in with the latest updates.

Tag: #homelab3 results

Gemma 4 Tops 45-Test Homelab LLM Benchmark

Gemma 4 Tops 45-Test Homelab LLM Benchmark

User benchmarked 19 local LLMs on AMD Strix Halo for homelab tasks like email classification, meal planning, and Home Assistant YAML. Gemma 4 26B-A4B ranked top after bug fixes, excelling in structured output and reliability. Custom 45-test suite scored by Claude Opus.

Reddit r/LocalLLaMACommunityApr 3#benchmarks#homelab#local-llms
🤖

Homelab Consolidates to 122B Qwen MoE

User benchmarks and consolidates homelab from three models to a single 122B Qwen3.5 MoE on Strix Halo, achieving 27.4 tok/s and top scores. IQ3 quantization matches Q4_K_M performance at half VRAM. Handles multiple apps like email, finance, and cameras concurrently.

Reddit r/LocalLLaMACommunityMar 27#homelab#moe#quantization