The new ChatGPT feature is now available at no cost.
OpenAI launched Canvas, a collaborative tool that allows users to preview how artificial intelligence will respond to their writing or programming requests.
OpenAI has announced the expansion of its Canvas feature, which debuted in October as a beta tool accessible only to Plus and Teams subscribers. From now on, Canvas will be available to all users, including those on the free tier. This collaborative interface allows users to visualize how artificial intelligence responds to writing or coding requests.
Despite appearances, Canvas is not simply a replica of Anthropic's Artifacts feature, as OpenAI is incorporating multiple innovations. Canvas is now natively integrated into the GPT-4o model within ChatGPT, eliminating the need to select it manually from the available models. Additionally, it can execute Python code within the interface, making it easier for ChatGPT to analyze and suggest improvements to the code.
Users can send both prose text and code directly to Canvas, avoiding the process of uploading content through the conventional chat interface. Furthermore, Canvas is now compatible with OpenAI's custom GPTs, simplifying their design and construction, and guiding those without deep programming knowledge. In case of errors, the 'Show Changes' feature helps highlight recent modifications in the text or code.
Canvas is accessible on the web and through the Windows desktop application. OpenAI has indicated that it will continue to make improvements and roll out new features in the near future. This update comes as part of its "12 Days of OpenAI" marketing campaign. Recently, the company also introduced Sora, its video generation system, which allows for the creation of 20-second clips in resolutions up to 1080p, and has launched a Pro subscription tier at $200 per month, necessary to access its range of reasoning models.
On another note, the French AI startup Mistral has introduced six new functionalities in its free assistant le Chat, which can now search for information on the web and provide citations, similar to other services like Perplexity and SearchGPT. It also includes a canvas feature for editing content and has integrated with Flux Pro from Black Forest Labs to generate images.
Finally, OpenAI is expected to launch a new AI agent known as Operator, which will allow ChatGPT to perform tasks on behalf of the user starting in January. This development, aimed at optimizing the user experience, will be implemented as part of the API for developers. Additionally, OpenAI has announced a new search feature, ChatGPT Search, which will provide real-time responses to user requests, posing a clear challenge to Google.
