OpenAI cuts GPT-5.6 Luna price 80 percent, rolls Fast mode into the API

OpenAI has sharply cut prices on two of its GPT-5.6 family models and replaced its Priority Processing add-on with a new Fast mode option across the API.

What's new

Per OpenAI's developer changelog, effective July 30, 2026: "Starting July 30, GPT-5.6 Luna costs 80% less, while GPT-5.6 Terra costs 20% less." The same entry describes the accompanying speed change: "We're also introducing Fast mode in the API, which replaces our Priority Processing offering. For GPT-5.6 Sol, Fast mode now delivers up to 2.5× faster speeds than standard processing at twice the price."

The transition is designed to be non-disruptive: OpenAI says the change is backward compatible, and "requests tagged priority will automatically use Fast mode" — existing integrations built around the old Priority Processing flag keep working without code changes, just under the new name and pricing structure.

GPT-5.6 Luna and GPT-5.6 Terra sit in OpenAI's current GPT-5.6 lineup alongside GPT-5.6 Sol; Luna and Terra are generally the lighter-weight, cheaper tiers relative to Sol, which is the model OpenAI recommends for production API workloads.

Context

This lands in the same stretch of weeks where OpenAI has been iterating quickly on the GPT-5.6 family's cost and speed knobs — the changelog shows Fast mode being extended to long-context prompts over 272K tokens just days earlier, and per-API-key usage tracking added around the same time. Price cuts on the smaller, cheaper models in a lineup are a familiar pattern industry-wide as compute costs fall and competitors like Google's Gemini Flash tier and open-weight models from labs such as DeepSeek and Qwen keep pressure on low-end API pricing.

Why it matters

An 80% price cut on GPT-5.6 Luna is a large move for a model that's presumably already positioned as OpenAI's budget option, and it puts more pressure on rival low-cost tiers to keep pace. Folding Priority Processing into a rebranded, backward-compatible Fast mode also simplifies OpenAI's pricing surface at a moment when the company has been adding several parallel speed and cost controls (fast mode, effort levels, priority tiers) to the API — collapsing two into one is a small but telling signal that the platform is trying to reduce its own configuration sprawl even as it keeps shipping new levers.