OpenAI's GPT-5.6 is now available worldwide, introducing three model tiers: Sol, Terra, and Luna. Each tier targets different performance and cost profiles, with Sol being the most capable and Luna the most economical. The API pricing structure has been updated to reflect these tiers, offering developers more flexibility in choosing the right balance of intelligence and cost. For indie hackers and startups, this means they can now opt for a cheaper model for simpler tasks while reserving the top-tier model for complex reasoning. The move is likely to intensify competition among AI providers, as pricing granularity becomes a key differentiator. Developers should evaluate their typical workloads and latency requirements before committing to a tier, as the cost differences can be substantial at scale.
OpenAI has launched GPT-5.6 globally with three distinct model tiers: Sol, Terra, and Luna. This article breaks down the differences in capability and API pricing, helping developers choose the right model for their use case. The tiered approach signals a shift toward more granular pricing, which could significantly impact how indie developers and enterprises budget for AI inference.