OpenAI has opened the API for GPT-5.5, the day after the model debuted in ChatGPT and Codex. The model ID is gpt-5.5, with snapshot version gpt-5.5-2026-04-23.
API pricing: $5 per million tokens for input, $0.50 per million tokens for cached input, and $30 per million tokens for output. The context window is 1.05 million tokens, with a maximum output of 128K tokens. Knowledge cutoff is December 2025. For inputs exceeding 272K tokens, input and output prices double and increase by 1.5x respectively. The model's efficiency — same latency, half the cost — was detailed in our earlier report.
GPT-5.5 Pro API is also live, with input at $30 and output at $180 per million tokens. Batch API and Flex pricing are available. Reasoning effort supports five levels: none, low, medium (default), high, and xhigh. After appearing in Codex last week without working, the API is now fully operational.