xAI is advancing its Grok platform with updates that reflect the needs of users who rely on specialized AI models for web-based research and productivity. The latest development introduces the ability to select Grok modes, such as expert, auto, or heavy, at the project level, which benefits users managing multiple workflows or seeking more precise outputs depending on their use case. The project-level selector makes it possible to standardize outputs for a team or individual projects, catering to professionals, researchers, and enthusiasts looking for different levels of detail or resource usage.
Behind the scenes, xAI is preparing more robust options for model and mode selection. Previously, Grok only allowed for switching between different processing modes, but unreleased changes point to a combined mode/model selector, where users could not only pick a mode but also choose the underlying model. This would be especially useful as new models, such as the anticipated Grok 4.20, become available. The option to revert to a previous version could be valuable for those who require consistency in results or need to replicate findings across different versions.
One notable feature under development is “background conversations.” This tool would allow users to start a long-running query in one chat and, while Grok processes the request, continue chatting or asking new questions within the same interface. For heavy research or iterative problem-solving, this could be a game-changer, as it reduces downtime and lets users multitask efficiently. The number of concurrent background processes has yet to be clarified, and there is no specific launch date. For xAI, these features fit into a broader strategy of making Grok a more versatile and competitive web AI solution, keeping pace with rapid developments in the industry and responding to the needs of technical and non-technical users alike.
