OpenAI slashes GPT-5.6 Luna price by 80%, undercutting Gemini and Claude

OpenAI announced major price cuts for GPT-5.6 models, with Luna dropping 80% to $0.20/M input tokens, making it cheaper than Google Gemini 3.1 Flash-Lite and Anthropic Claude Haiku 4.5. The reduction was enabled by GPT-5.6 Sol optimizing inference kernels.

OpenAI reduced prices for GPT-5.6 models on July 30, 2026: GPT-5.6 Terra got a 20% cut, while GPT-5.6 Luna dropped 80% to $0.20 per million input tokens and $1.20 per million output tokens. This makes Luna cheaper than Google's Gemini 3.1 Flash-Lite ($0.25/$1.50) and far below Anthropic's Claude Haiku 4.5 ($1/$5). The price reduction was achieved using GPT-5.6 Sol, which optimized load balancing and inference kernels. GPT-5.6 Sol rewrote production kernels in Triton and Gluon, reducing serving costs by 20%. Simon Willison, the author, switched his demo site agent.datasette.io from Gemini 3.1 Flash-Lite to Luna.

Source: Simon Willison — original