Elon Musk took a moment on August 25 to publicly endorse a review of Grok, tagging the newly launched Grok Bot account in the process. It's a small signal, but it lands in the middle of a genuinely busy stretch for SpaceXAI — new model versions, a fresh agent product, a corporate rebrand, and at least one public admission that the AI isn't winning every race. Here's what's actually worth knowing right now.
1. Grok Bot is the product Musk is actually promoting
When Musk wrote "Good review of Grok @Bot," the tag wasn't decorative. Grok Bot — an early-beta AI agent running on a persistent cloud computer — launched on August 11, 2026, and expanded to SuperGrok Plus, SuperGrok Heavy, Cursor Pro+, Cursor Ultra, and Cursor Teams plans by August 21, according to SpaceXAI. The endorsement is effectively a nudge toward the agent tier, not the base chatbot most casual users know.
2. The underlying model just got a significant upgrade
Grok 4.6 launched officially on August 12, 2026, with a 500,000-token context window and an API model ID of grok-4.6. SpaceXAI positions it as the flagship for coding, agentic tasks, and knowledge work. API pricing sits at $2 per million input tokens and $6 per million output tokens — rates that double for prompts exceeding 200,000 tokens. A further update, Grok 4.7, is anticipated before September 2, according to published roadmap signals.
3. The corporate structure has changed — this is no longer xAI
SpaceX acquired xAI in February 2026, and the combined entity was formally renamed SpaceXAI earlier in August 2026. That matters for context: when Musk promotes Grok, he's promoting a product now embedded inside a much larger aerospace and technology conglomerate. The Cursor acquisition — reportedly valued at $60 billion — also folded into this structure, making Grok Bot's tight integration with Cursor plans less of a partnership and more of an internal product decision.
4. Musk has simultaneously acknowledged Grok is behind
The endorsement tweet exists alongside a more candid admission. According to reporting, Musk told employees at an all-hands meeting that Grok is "losing the AI race to Anthropic" and has "fallen behind competitors," adding that he is "not used to losing" and committed to building the largest computing capacity in the industry. That tension — public boosterism alongside private urgency — is the actual story underneath a three-word tweet.
5. There was a glitch the day before the endorsement
On August 24, 2026 — one day before Musk's post — some Grok users reported receiving bizarre and nonsensical responses from the chatbot. SpaceXAI attributed the issue to a "rare temporary generation glitch," with earlier signs of the problem appearing in Grok Lite as far back as August 19. The timing is worth noting: a high-profile founder endorsement following a visible reliability stumble is a classic reputation-management move, whether intentional or not.
The broader picture is of an AI product moving fast — new models, a new agent tier, a new corporate home — while still chasing the benchmark leaders. Musk's three-word post is easy to scroll past, but the context around it tells a more complicated story about where Grok actually stands heading into September.
Sources & reporting notes
The links below identify the material source records used for this report.
- @elonmusk on X (2026-08-25T21:49:27.000Z) — Direct source
Source links are preserved as published or accessed. See our editorial standards and corrections policy.
The BASENOR Editorial Desk covers Tesla, SpaceX, and related technology, curating reporting from primary sources — official accounts, regulatory filings, and software release data. Every article passes source-record and fact-checking review before publication. About the newsroom.
This report was curated by the BASENOR Editorial Desk from the sources listed above. Read our editorial standards or email editorial@basenor.com to report an error.
