Parked: Tier 5 Research

Tier 5 research is paused:

• reasoning on 22M
• test-time compute scaling
• vision-language toy
• diffusion LM micro
• sparse MoE kernels
• TTS toy
• explainer-video model

These are learning-rich but not the current factory proof.

Reopen after the factory can repeatedly produce, evaluate, package, and report a specialist.