← AI PulseAug 20, 2026

Deep · news · Multi-source brief

xAI Introduces Multi-Image Editing and Realtime Multi-agent Research

xAI has introduced multi-image editing capabilities for its Grok Imagine model, allowing users to combine up to three source images for a single edit, alongside a beta release of Realtime Multi-agent Research for Grok 4.20-multi-agent.

By Illumora Editorial

Source · Aug 20, 2026, 3:13 PM · On Illumora · Aug 20, 2026, 4:09 PM

Media from the primary source — shown here so you can stay on Illumora.

Synthesized from multiple allowlisted primaries on the same event. Lanes →

Brief drafted by Illumora’s editorial model from the linked primary source. Ops desk reviews flagged pieces. How we write →

Read the source →xAI Docs — Multi-Image Editing | SpaceXAI Docs
Save

xAI has expanded the capabilities of its Grok models, according to its developer documentation. The Grok Imagine model now supports multi-image editing, enabling the use of multiple source images for a single edit. Additionally, xAI has released Realtime Multi-agent Research in beta, a feature designed for orchestrating multiple AI agents to perform complex research tasks.

Key Points

  • The Grok Imagine model supports multi-image editing, allowing up to three source images for a single edit.
  • Users can specify the order of images in the request, and the output aspect ratio defaults to the first input image.
  • The aspect ratio can be overridden using a parameter, such as "1:1" or "16:9".
  • Source images can be public URLs, base64-encoded data URIs, or file_ids from the Files API, with mixed types allowed in a single request.
  • Realtime Multi-agent Research is available in beta and uses the grok-4.20-multi-agent model.
  • This multi-agent feature orchestrates specialized AI agents for deep, multi-step research tasks.
  • The API interface and behavior for Realtime Multi-agent Research are subject to change as it is in beta.

Context

According to xAI's documentation, the multi-image editing feature for Grok Imagine provides flexibility in how source images are provided, supporting various formats and combinations within a single request. The documentation also states that Realtime Multi-agent Research allows Grok to coordinate multiple AI agents that specialize in different aspects of research, such as searching the web, analyzing data, and synthesizing findings. This collaboration aims to deliver comprehensive, well-sourced answers, moving beyond single-turn tool use.

Why It Matters

These updates offer builders new ways to integrate advanced image manipulation and complex research orchestration into their applications. The multi-image editing feature provides granular control over image inputs, while the multi-agent research capability introduces a framework for more sophisticated, multi-step AI-driven workflows, though its beta status indicates potential for breaking changes.

What To Do

  • Review the Grok Imagine documentation for details on multi-image editing parameters and supported image input types.
  • Explore the Files API Integration section for examples of using file_ids with Grok Imagine.
  • For multi-agent research, note that the grok-4.20-multi-agent model is currently in beta and its API interface may change.
  • Test the Realtime Multi-agent Research feature with the grok-4.20-multi-agent model to understand its current capabilities and limitations.

Keep Exploring

/atlas/**grok**-family