Grok Imagine upgrades bring voice consistency, native 1080p video, and multi-reference support to xAI’s creative toolkit

XAI rolled out an upgrade to Grok Imagine on August 1, adding multi-channel audio references for voice consistency, native 1080p video output, and support for up to seven image references, available to paid subscribers. The update, referred to as Grok Imagine Video 1.5 or v1.3.28, aims to improve character and scene stability in AI-generated content, with free-tier access reduced as part of xAI's monetization strategy.

Via xai-technologies.com Grok Imagine upgrades bring voice consistency, native 1080p video, and multi-reference support to xAI’s creative toolkit xAI's latest update to its AI image and video generator adds up to three audio inputs, seven image references, and full HD output for paid subscribers. xAI just dropped a meaningful upgrade to Grok Imagine, the AI-powered image and video generation tool baked into the Grok chatbot on X. The update, rolled out on August 1, introduces improved voice consistency through multi-channel audio references, native 1080p video output, and expanded multi-reference support for maintaining character and scene stability across generated content. In plain terms: creators can now feed the system up to three audio inputs and seven image references simultaneously, and get back high-definition video that actually looks like it was meant to exist. No janky upscaling required. What actually changed The headline feature is voice consistency. Previous versions of Grok Imagine had built in lip-sync and audio generation capabilities, but keeping a character’s voice stable across scenes was, let’s say, a work in progress. The new multi-channel audio reference system is designed to fix that, allowing creators to anchor their generated content to specific voice profiles rather than hoping the AI doesn’t suddenly give their protagonist a different accent mid-scene. The multi-reference system is the other big deal here. Supporting up to seven image references means creators can lock down visual consistency for characters, environments, and objects across multiple generations. Then there’s native 1080p output for both text-to-video and image-to-video pipelines. This sounds incremental until you realize that most competing AI video tools still output at lower resolutions and rely on post-processing upscaling to hit full HD. Eliminating that step isn’t just a quality improvement, it’s a workflow improvement. The update has been referred to as Grok Imagine Video 1.5 or v1.3.28, building on xAI’s Aurora-driven models that have been iterating rapidly over the past year. The subscriber-only play Here’s the thing about all these shiny new features: they’re behind a paywall. Free-tier users on X have seen their access to image and video generation pulled back significantly in 2026. The upgraded capabilities are subscriber-exclusive, which is a deliberate monetization bet by xAI. Elon Musk and xAI have signaled that iterative improvements will continue at a regular cadence, with community-driven techniques for video extension and prompt-based consistency controls already in the pipeline. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .