M5 Max Glossary

Apple Silicon system-on-chip, M5 generation, Max-tier configuration. The operator's primary workstation; the substrate for local model inference and the *Sovereign Compute* setup.

The M5 Max is the highest-tier laptop-class system-on-chip in Apple’s M5 generation, released in 2025–2026. The operator of this Dictionary purchased a MacBook Pro 16” M5 Max configuration in April 2026: 18-core CPU, 40-core GPU, 16-core Neural Engine, 128 GB unified memory, 4 TB SSD. The order was placed April 18, 2026 (order W1429262259) for $5,890.73 plus AppleCare One, with delivery in early May 2026. See the Sovereign Compute entry and hardware/purchases/2026-04-18_macbook_pro_m5_max.md in the workspace for the full purchase rationale.

For this Dictionary, the M5 Max is the substrate of the local-compute argument. The 128 GB of unified memory and the 40-core GPU are sufficient to run open-weights models in the 30B-active-parameter range (Qwen 3.6 30B A3B, Gemma 4 31B, Hermes derivatives in the 70B range with quantisation) at usable speeds for sustained operator work. The architecture is documented in TOOLS.md: hybrid dense+MoE routing, model-to-task pairing, and the incremental-construction workflow for complex local tasks.

The structural significance is named in GenXClaw and FERPA Compliance Posture: the M5 Max is, accidentally, the architecture FERPA law would have specified if FERPA had been written with student-AI workflows in mind. The hardware purchase predates the FERPA argument; the convergence is real.

See also

Return to Dictionary All Entries (A–Z) For Students Other Writing