> For the complete documentation index, see [llms.txt](https://sincronix.gitbook.io/sincronix/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://sincronix.gitbook.io/sincronix/features/gpt-4-introduction-test.md).

# GPT-4 Introduction (Test)

Under test phase - GPT 3 ready for use

By introducing GPT-4, a large multimodal model capable of accepting text and image inputs and providing text outputs, SincroniX has integrated deep learning technology into its system. The performance of GPT-4 has been impressive, demonstrating more reliability, creativity, and nuanced instruction handling than its predecessor, GPT-3.5, although it may not outperform humans in all real-world scenarios.

GPT-4's ability to accept both text and image inputs enables a wide range of language and vision tasks, generating output in natural language, code, and other forms using mixed text and image inputs. GPT-4 displays similar capabilities in various domains, including documents containing text, photographs, diagrams, or screenshots, as well as in text-only inputs. In addition, the performance of GPT-4 can be improved through test-time techniques such as few-trigger and thought-chain prompts, which were initially designed for text-only language models. However, image inputs are currently in the preview phase of the trial and are not yet available to the audience.