OpenAI: GPT-5.6 Luna

OpenAI's GPT-5.6 Luna, created on July 9, 2026, is positioned as a fast, cost-efficient model within the GPT-5.6 series, designed for high-volume, latency-sensitive tasks. It consistently performs among the fastest models, ranking in the 84th percentile across eight benchmarks, and offers competitive pricing, placing in the 40th percentile. Notably, GPT-5.6 Luna demonstrates exceptional reliability with a 100% success rate across all benchmarks, indicating minimal technical failures. The model exhibits strong performance across various categories. It achieves perfect accuracy in General Knowledge, making it the most accurate model at its price point and among models of comparable speed. In Coding, it boasts 95.0% accuracy, also standing out as the most accurate among models with similar speed. Reasoning and Mathematics benchmarks show robust performance with 90.0% and 94.0% accuracy respectively. While its Hallucinations accuracy is 98.0%, indicating a good ability to acknowledge uncertainty, there's a slight room for improvement. Instruction Following is solid at 81.0%, and Email Classification is strong at 98.0%. Ethics performance is also high at 99.0%. Overall, GPT-5.6 Luna's key strengths lie in its speed, reliability, and high accuracy in knowledge-based and coding tasks, making it well-suited for its intended applications.