OpenAI Slashes GPT-5.6 API Prices by Up to 80%

New API Costs for GPT-5.6 Luna

GPT-5.6 Luna — the fastest and most affordable model in the lineup — saw an 80% price cut . Its new API pricing is :

This brings Luna's combined input-plus-output price to $1.40 per million tokens . For context, the mid-tier GPT-5.6 Terra was reduced by 20% to $2.00 per million input tokens and $12.00 per million output tokens (down from $2.50 and $15.00, respectively) . The flagship GPT-5.6 Sol remained at $5.00 input / $30.00 output per million tokens .

The new prices took effect immediately on July 30, 2026, and apply to the standard API tier .

Stated Reasons for the Price Cuts

OpenAI provided two interrelated explanations for the aggressive price reductions .

1. Efficiency Gains from GPT-5.6 Architecture

In its official blog post, OpenAI stated the cuts were enabled by improvements across every layer of the model, including better code efficiency and system optimizations that allow it to deliver "substantially more intelligence per dollar" . The company framed this as "advancing the price-performance frontier" . Notably, efficiency breakthroughs were partly driven by GPT-5.6 Sol itself, which autonomously rewrote production code kernels, cutting model serving costs by 20% . Additional kernel-level work reduced end-to-end serving costs by 20%, while experiments increased token-generation efficiency by more than 15% .

2. Competitive Pressure and Enterprise Cost Sensitivity

Multiple reports note the pricing move came amid growing enterprise scrutiny of AI spending and intensifying competition from cheaper alternatives, including Chinese open-weight models . CNBC reported that OpenAI is "facing pressure to cater to a more cost-sensitive customer base" where enterprises have been hesitant to deploy expensive AI models without a clear picture of return on investment . Axios noted that "cheaper Chinese open-weight models have also increased pressure on OpenAI and Anthropic to prove that their models justify their higher costs" . Reuters similarly observed that the cuts were "partly enabled by efficiency gains from GPT-5.6" but also reflected that "businesses scrutinize AI spend" .

CEO Sam Altman announced the changes on X (formerly Twitter), writing: "We want to offer the best price/intelligence tradeoff at every level" .

What This Means for Developers

For developers and enterprises using the OpenAI API, the 80% reduction in Luna's price makes it one of the most cost-effective options for high-volume workloads. Luna is positioned as the fastest and most affordable model in the GPT-5.6 series, ideal for tasks requiring speed and efficiency . The price cuts also apply to how usage is counted against paid subscriptions in ChatGPT Work and Codex, effectively increasing the amount of AI work enterprise subscribers can perform without paying more .

In summary, OpenAI's July 30, 2026 pricing update represents a sharp recalibration of its API costs, driven by both genuine engineering progress and the realities of an increasingly competitive AI landscape.