[tag]

#local-llm

3 stories

[tooling]Aug 20, 2026By ByteBulletin Editor

local_bench: A zero-config TUI to benchmark the LLMs already on your machine

A new open-source tool turns your local model collection into a live terminal leaderboard measuring speed, memory, energy, and quality in one command.

[models]Aug 9, 2026By ByteBulletin Editor

Escha Labs ships 2-bit Qwen3.6-35B-A3B build that runs on a 24GB GPU

A 12.3GB MoE quant packs a 35B model onto consumer cards, with quality within noise of FP8 on most benchmarks.

[research]Aug 8, 2026By ByteBulletin Editor

ExpertCache lets a 120B Mixture-of-Experts model run on a 16GB Mac — slowly, but surprisingly intact

Amos Labs' experimental Metal runtime proves oversized sparse MoE checkpoints can fit in constrained Apple Silicon memory without sacrificing functional capability, even if interactive speeds remain out of reach.