Show HN: Gemma 4 26B at 100 tok/s on a Mac, using the Neural Engine and 2GB RAM
Description and use
Gemma 4 26B runs on a Mac at 100 tokens/second using the Neural Engine and 2 GB RAM, making it suitable for prototypes and internal workflows. Start by manually inspecting compatibility, updates, and documentation, then test the model in a small project. With regular maintenance and dependency checks, it is considered production‑ready for moderate workloads.
Use in a workflow
Show HN: Gemma 4 26B at 100 tok/s on a Mac, using the Neural Engine and 2GB RAM may be useful when its README and activity match a concrete workflow.
Limitations and review notes
Medium: useful for prototypes or internal workflows, with dependency and maintenance checks before production.
- Source: hackernews
- Record date: 2026-10-21T01:34:57.000Z