OpenAI cuts GPT-5.6 Luna price 80% and Terra price 20%

OpenAI slashed API pricing for two of its three GPT-5.6 models on July 30, cutting the cost of GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20%, according to the company's official API changelog.

What's new

The changelog entry states plainly: "Starting July 30, GPT-5.6 Luna costs 80% less, while GPT-5.6 Terra costs 20% less." The cuts apply to both the Responses API (v1/responses) and Chat Completions (v1/chat/completions) endpoints, and OpenAI points developers to its pricing page for the updated per-token rates.

Luna is the smallest, most efficient tier in the GPT-5.6 family, built for high-volume workloads, while Terra sits in the middle as a balance of intelligence and cost. Sol, the flagship frontier tier, is not part of this price cut.

The move follows reporting from CNBC that OpenAI is responding to growing enterprise cost-sensitivity as companies scale up AI usage and scrutinize their inference bills more closely.

Context

OpenAI launched the GPT-5.6 family — Sol, Terra, and Luna — on July 9, with the plain gpt-5.6 alias routing to Sol. Since then, the company has continued to iterate on the lineup: it has published research showing Sol's ARC-AGI-3 benchmark score tripled after enabling two existing API settings, and it has steadily added infrastructure features like Terraform support and hard spend limits for organizations managing API budgets.

Aggressive price cuts on lower-tier models are a familiar pattern in the frontier-lab race. As open-weight competitors like DeepSeek and Moonshot's Kimi K3 undercut US labs on price-per-task — a recent Together AI benchmark found Kimi K3 delivers 2.8x more solved tasks per dollar than GPT-5.6 Sol — OpenAI has incentive to keep its cheaper tiers competitive on cost even as it holds the line on Sol's premium pricing.

Why it matters

For developers building high-volume applications — customer support bots, batch classification, content moderation — Luna's 80% price cut meaningfully changes the unit economics of running GPT-5.6 at scale. Terra's smaller 20% cut nudges OpenAI's mid-tier model closer to price parity with rivals occupying the same intelligence-cost middle ground, such as Anthropic's smaller Claude models or Google's Gemini Flash line.

The timing — three weeks after GPT-5.6's launch — also signals that OpenAI is willing to move quickly on pricing once a model family is in market, rather than waiting for a full generational refresh. That's a meaningful signal for enterprise buyers weighing multi-model strategies: prices on frontier-adjacent tiers may keep falling faster than in previous OpenAI release cycles, which changes the calculus on locking into long-term contracts at current rates.