OpenAI’s GPT-5.6 trio just landed in Pickaxe, giving you more control over the trade-offs between reasoning ceiling, speed, and cost. Here’s what each one brings to the table:
Sol
• Highest reasoning ceiling in the family
• Best used when marginal reliability matters like critical migrations, security reviews, and long-running agentic flows where retries are expensive
• The priciest of the three - best for users with intensive workflows who need the extra compute power
Terra
• Balanced default for everyday interactive coding and agentic work
• Reasoning ceiling that is effectively tied with Anthropic’s Claude Fable 5
• More modestly priced, and great to use when Luna runs out of steam but Sol would be overkill
Luna
• Lightweight, cost-efficient option for smaller or high-volume tasks
• Performs highly on multiple coding benchmarks
• Lowest pricing of the group - ideal as the default routing for bulk coding, copy, and support automation
Give them a try and let us know what you think! Happy prompting 
