OpenAI has launched GPT-5.5, a new large language model featuring increased capabilities across a broad variety of categories. The model is significantly better than its predecessors at solving math problems and writing code. It is available in two versions: a standard edition and a more capable, significantly pricier edition.
Codenamed Spud, the model is designed to complete complex multi-step tasks with minimal human direction. It sets new benchmarks in knowledge work, computer use, and agentic coding, while matching the per-token latency of GPT-5.4. According to OpenAI, the model also offers deeper research capabilities.
The release follows the launch of a new LLM from rival Anthropic PBC a week prior, as OpenAI seeks to keep pace with competitors in courting business customers. While the maker of ChatGPT is taking a more open approach to cybersecurity than Anthropic, API access for the new model is delayed pending additional safety work.