⚡️ Introducing GPT‑5.4
This model is now available in two versions – Thinking and Pro. In coding tests, it shows modest achievements; in mathematics, quite respectable results; and especially in tasks related to computer usage, it excels. The latter are particularly proud of.
Additionally, GPT‑5.4 has become more efficient: it uses fewer tokens for reasoning and responds faster. However, one point to note is that its price has slightly increased.
What else is interesting:
– Now there is an option to stop the model during response and provide additional instructions to adjust its behavior.
– The context window has been expanded to a full one million tokens — finally! This significantly enhances the ability to work with large volumes of data.
– Codex has introduced an accelerated mode /fast, which allows generating responses approximately 1.5 times faster. The same model with the same reasoning capabilities is used; only the speed has increased. However, keep in mind that in this mode, each token counts twice toward the limits.
More information can be found on the official OpenAI website.
Created with n8n:
https://cutt.ly/n8n
Created with syllaby:
https://cutt.ly/syllaby