Which model to pick? Opus 5 — Coding / Claude Code / Agents Fable 5 — Advanced Reasoning / Complex Coding / Research GPT-5.6 — All-Rounder / API / Agents / Automation GPT-6 Astra — Extreme Reasoning / Terminal / Heavy-Duty Agents Opus 4.8 — Reliable Coding / Claude Code Not sure which one to choose? General use → GPT-5.6 Coding → Opus 5 Hardest problems → Fable 5 / Astra Claude Code → Opus 5 Long-running Agent tasks → Opus 5 / Fable 5 / Astra