I ran Gemma 4 as a local model in Codex CLI

Based on 5 sampled comments from this thread. Analyzed 2026-04-13.

• Praise: Running Gemma 4 on local hardware like the M3 Ultra demonstrates the model's capability to handle large context sizes effectively, which is crucial for complex tasks.
• Praise: The integration of Gemma 4 with tools like Zed via ACP showcases its versatility and ease of use in real-world applications.
• Praise: Users report significant performance improvements with the M5 Pro, indicating that hardware upgrades can lead to enhanced model efficiency and speed.
• Praise: The ability to offload MoE tasks to the CPU allows for flexibility in managing context size and token generation speed, which can optimize performance.
• Concern: Concerns about the quality of fine-tuned models compared to original versions suggest skepticism about performance claims, highlighting the need for thorough benchmarking.
• Concern: Issues with tool calling in models like Qwen 3.5 indicate that practical challenges remain, which can hinder user experience and effectiveness in coding tasks.
• Concern: The necessity of increasing context size to achieve useful results points to potential limitations in the model's default configurations, which may not meet all user needs.