Llama 4: The New Open-Weights Baseline

By C.W. Jameson · Published 20 October 2025 · Last reviewed 20 October 2025

Meta shipped a model that can be self-hosted on eight H100s and matches GPT-4o on most benchmarks. The inference industry responded within two weeks.

An analysis of Llama 4's MoE architecture, benchmark performance, and what it means for the open-weights ecosystem.