OpenAI today announced the introduction of the GPT-5.4 model, allowing users to use it on ChatGPT through the GPT-5.4 Thinking option, in addition to access to the API and Codex.
OpenAI states that this model is the most knowledgeable model especially for professional work, and is tailored for reasoning, programming and agent-based AI work. The model also offers support for up to 1 million tokens for context.
The new model is also said to be able to complete a number of tasks using a lower token rate than the previous version. In addition, the hallucination rate is also reduced, and it is said to bring a 33% reduction in making errors compared to the previous GPT 5.2 version.
For those of you who use ChatGPT, you can try GPT-5.4 Thinking on ChatGPT today.
