OpenAI Revolutionizes AI Costs with GPT-5.6 Luna and Terra

⚡

Key Takeaways

1OpenAI reduces the price of GPT-5.6 Luna by 80%, making AI more accessible for businesses.

2The GPT-5.6 Terra model sees its cost decrease by 20%, optimizing daily work with increased efficiency.

3The new Fast mode in the API accelerates the processing of GPT-5.6 Sol by up to 2.5 times, without loss of intelligence.

💡Why it matters — These cost reductions and speed improvements enable businesses to integrate AI more cost-effectively and efficiently into their operations.

OpenAI Transforms AI Efficiency with GPT-5.6

OpenAI has recently announced significant improvements in the efficiency of its artificial intelligence models, aiming to deliver optimal performance at a reduced cost for businesses. These advancements specifically pertain to the GPT-5.6 model, which comes in several versions tailored to various professional needs.

Cost Reductions for Luna and Terra

Starting today, the GPT-5.6 Luna model, known for its speed and affordability, sees its price reduced by 80%. Meanwhile, GPT-5.6 Terra, designed for balanced daily use, benefits from a 20% price cut. These pricing adjustments also apply to the use of Codex and ChatGPT Work services, allowing companies to manage significant workloads more economically while maintaining high quality.

A Commitment to AI Accessibility

OpenAI's mission is to make advanced artificial intelligence accessible and beneficial for all of humanity. These price reductions illustrate this commitment, the result of several years of optimizing the processes of building, deploying, and utilizing models.

Introduction of Fast Mode

OpenAI has also introduced the Fast mode in its API, replacing the previous Priority Processing offering. For the GPT-5.6 Sol model, this mode allows processing up to 2.5 times faster than the standard, at double the cost, while maintaining the same level of intelligence. This mode is backward compatible, meaning that requests marked as priority will automatically benefit from this acceleration.

User Feedback

Michele Catasta, President and Head of AI at Replit, praised GPT-5.6 Luna as the most affordable and powerful model, opening new possibilities for Replit.
Hoda Noorian, Head of AI Product at Notion, found GPT-5.6 Terra ideal for daily tasks, offering quality comparable to GPT-5.5 at reduced cost and time.
Shaiyon Hariri, creator of Ramp SWE-Bench, noted the importance of the GPT-5.6 series in maximizing results per dollar.
Sid Pardeshi, CTO and co-founder of Blitzy, described GPT-5.6 Luna as a major advancement in agentic behavior since GPT-4o mini.

Optimizing Outcomes

Effective use of AI begins with defining desired outcomes. GPT-5.6 provides businesses with the flexibility to optimize the balance between intelligence, speed, reliability, and cost. The Luna model, for example, delivers performance comparable to last year's leading models for about 6 cents per task, with speed increased by nearly nine times.

Efficiency Advances

The efficiency improvements in GPT-5.6 stem from the redesign of models, inference systems, and integration with tools and context. These models take a more direct path, reducing time, tokens, and cost per outcome. GPT-5.6 Sol plays a key role in optimizing processes, rewriting and enhancing production cores, and conducting experiments to refine token generation.

Large-Scale Computing Strategy

To meet the growing demand for artificial intelligence, OpenAI is developing a resilient infrastructure, matching each workload with the most suitable systems. This strategy supports the optimization of the price-performance curve.

Availability and Pricing

The GPT-5.6 Terra and Luna models are available through ChatGPT Work, Codex, and the OpenAI API. Starting July 30, API rates are set at $2 per million input tokens and $12 per million output tokens for Terra, and $0.20 per million input tokens and $1.20 per million output tokens for Luna. Prices for Sol remain unchanged.