Last week, @X sued several people who abused Creator Revenue Sharing by operating a coordinated network of accounts, posting inauthentic content to manipulate engagement, and using multiple bank accounts to hide their scheme. We do not tolerate fraudulent behavior on X -- and will act forcefully to protect our platform and the earnings of genuine creators. You can read our lawsuit here:

Grok Build v1.0.39 is out with a few practical updates. Standouts include subagents can always use the parent model, with that choice locked in when. Earlier image attachments now survive multiple compactions via a persisted path list. Also worth noting: subagent spawning no longer requires choosing a type. Release Notes: v1.0.39 - 2026-09-20 Features - Subagents can always use the parent model, with that choice locked in when the session starts. - Earlier image attachments now survive multiple compactions via a persisted path list. - Subagent spawning no longer requires choosing a type; omitted calls default to general-purpose. - The agent now keeps helper scripts, logs, and PR drafts in the system temp directory instead of the repository. - read_file descriptions now tell the model when offset/limit are ignored for SKILL.md and instruction files. - Effort levels for models now come from the API instead of hard-coded lists, and config aliases inherit the menu. - MCP tools that receive the wrong JSON shape are now automatically fixed when the mismatch is unambiguous. Bug Fixes - Local worktree capture no longer refuses repositories that contain uninitialized submodules. - Memory captures no longer include opaque model reasoning blobs. - Idle timeouts now show a clear message instead of raw internal text. - Images attached to the last user prompt now survive compaction. - Custom agent profiles now persist correctly across session resume and reload. - Subagent labels now appear as "Subagent" instead of "General" in transcripts and the tasks pane. Listen to the video below for more details.

Grok Build v1.0.38 is out. Long agent replies in prompt-suggestion transcripts are no longer cut off. Read_file on skill and instruction files can now be configured per deployment. Pasted images survive yank/undo/history recall and the user is notified when any. Release Notes: v1.0.38 - 2026-09-19 Features - Long agent replies in prompt-suggestion transcripts are no longer cut off mid-sentence. - read_file on skill and instruction files can now be configured per deployment. - Pasted images survive yank/undo/history recall and the user is notified when any cannot be sent. Bug Fixes - Subagent overlays no longer stay stuck showing "Cancelling" after the child turn ends. - Long quoted arguments in permission prompts now wrap instead of clipping. - Fixed tool-calling errors by sending inline-only use_tool schema. Listen to the video below for more details.

X clawed back the new X numbers feature yesterday and are planning to reintroduce it with easier to remember 8 digit numbers . The previous iteration seemingly made people think it was an actual 10 digit phone number they would be able to use. this is not the case. It's more like a password people can use to DM you when your inbox is locked down. The best part is you can share links that will let people message you directly until you rotate your X number . Ie : Use my 𝕏 Number 81116789 to chat with me:

Grok Build v1.0.37 is out. Mermaid flowcharts with complex labels or large sizes now render and open correctly. `/context` legend now shows accurate percentages that always add up to 100%. Pressing `y` on a queue row now copies the prompt text to the clipboard. Release Notes: v1.0.37 - 2026-09-18 Bug Fixes - Mermaid flowcharts with complex labels or large sizes now render and open correctly. - `/context` legend now shows accurate percentages that always add up to 100%. - Pressing `y` on a queue row now copies the prompt text to the clipboard. - `/model` effort picker now defaults to the model's recommended effort level. - Bash commands using quoted filename variables now respect configured permission rules instead of always prompting. Listen to the video below for more details.