# OpenAI sets GPT-5.6 pricing at $5 input, $30 output per 1M tokens with three-tier model family > Source: > Published: 2026-07-09 17:59:58+00:00 # OpenAI sets GPT-5.6 pricing at $5 input, $30 output per 1M tokens with three-tier model family The Sol, Terra, and Luna lineup gives developers a pricing menu that ranges from budget-friendly to flagship, with implications for AI-driven crypto infrastructure OpenAI just rolled out its GPT-5.6 model family with a tiered pricing structure. The flagship Sol model costs $5 per million input tokens and $30 per million output tokens, while cheaper options scale down for developers who don’t need the full horsepower. ## Three tiers, three price points The GPT-5.6 family breaks down into Sol, Terra, and Luna. Sol sits at the top with that $5/$30 pricing split for input and output respectively. Terra occupies the middle ground at $2.50 per million input tokens and $15 per million output tokens. Luna comes in as the budget option at $1 per million input tokens and $6 per million output tokens. The rollout followed a limited preview that began on June 26, 2026. Public availability kicked off on July 9, 2026, after a government review of the models’ capabilities. OpenAI CEO Sam Altman confirmed that the new models deliver a 54% increase in token efficiency on specific tasks compared to previous iterations. Enhanced prompt caching rounds out the technical upgrades. OpenAI is offering a 90% discount on cached reads along with a 1.25x input rate for cache writes. ## Why crypto builders should care OpenAI didn’t mention crypto, blockchain, or any token-related entities in its announcements. The tiered structure enables intelligent routing, where you send simple queries to Luna, moderate tasks to Terra, and only escalate complex reasoning to Sol. A crypto analytics platform could use Luna for basic transaction categorization, Terra for pattern recognition across wallet clusters, and Sol for sophisticated threat detection or novel attack vector analysis. The prompt caching feature deserves particular attention from crypto developers. A 90% discount on cached reads means that once a model has ingested a protocol’s documentation or a contract’s structure, subsequent queries about that protocol cost a fraction of the initial analysis. ## Competitive pressure and market implications Anthropic, Google, and a growing roster of open-source model providers are all competing for the same developer wallets. The three-tier approach is a direct response to that pressure. The 54% token efficiency improvement Altman cited compounds with the lower per-token prices, meaning the effective cost reduction for many workloads could be substantially larger than the sticker price suggests. **Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).