OpenAI shook up the AI pricing landscape by cutting the cost of its GPT-5.6 Luna model by a staggering 80%, dropping the price to just $1.40 per million combined input and output tokens. This move pushes Luna well below Google's Gemini 3.5 Flash-Lite, priced at $2.80, and far undercuts Gemini 3.6 Flash which stands at $9. The company also trimmed GPT-5.6 Terra’s price by 20%, now set at $14 per million tokens, matching Google’s Gemini 3.1 Pro Preview for larger context windows.
CEO Sam Altman announced these changes on X, calling them “major price cuts today,” just weeks after the GPT-5.6 series went broadly public. The timing is telling, arriving shortly after competitors Google and Anthropic unveiled their own pricing adjustments in a clear battle for cost efficiency in the AI market.
The Luna model’s cost drop from $7 to $1.40 per million tokens dramatically lowers barriers for developers processing millions of requests daily. While it doesn’t claim the absolute lowest price some providers like Xiaomi and MiniMax still offer cheaper rates this marks the first time an OpenAI frontier-series model has aggressively targeted low-cost tiers, positioning it alongside Google, Xiaomi, DeepSeek, and MiniMax in this competitive space.
OpenAI also introduced a premium Fast mode for its GPT-5.6 Sol model, offering up to 2.5 times higher throughput at $70 per million tokens, doubled compared to the Standard rate, catering to latency-sensitive workloads.
Anthropic’s Claude Opus 5 holds firm at $30 per million tokens, maintaining near-Fable 5 performance at the same price as its previous version, Opus 4.8.
This price recalibration could reshape AI adoption economics, especially for startups and high-volume users seeking affordable yet advanced models.
This material is informational and does not constitute financial advice.
