# Musk wrote that "@SpaceX will use the best back end model for any given task, including Claude Opus 5.5." Cursor's help article, undated, describes that routing as how Grok Bot works; the changelog has no entry for it, and the docs differ on whether users can see which model answered. Of 13 headlines the desk rated against the text, 7 are supported, 5 stretched and 1 unsupported.

- Canonical: https://thestochasticparrot.com/audits/grokbot-claude-routing
- Structured (spans, splits, sources): https://thestochasticparrot.com/audits/grokbot-claude-routing.json
- Date: 2026-10-10T04:09:43Z
- Context: An editorial: a longer-form argument from the desk, grounded across 38 sources — synthesis, not a single day's coverage.
- License: CC BY 4.0

---

[ EDITORIAL // grokbot-claude-routing // 2026-10-09 ]

# Grok Bot and Claude: Musk Said ‘Will,’ Coverage Said ‘Now,’ the Docs Name No Date

## Musk wrote that "@SpaceX will use the best back end model for any given task, including Claude Opus 5.5." Cursor's help article, undated, describes that routing as how Grok Bot works; the changelog has no entry for it, and the docs differ on whether users can see which model answered. Of 13 headlines the desk rated against the text, 7 are supported, 5 stretched and 1 unsupported.

The sentence at the center of this story names its actor, and the actor is not the product. It is @SpaceX.

── COMPATIBLE: the_original_post · what Musk wrote, in full ──
Elon Musk on X (post of 7 October 2026): "Important note regarding Grok @Bot:"
Elon Musk on X (post of 7 October 2026): "Going forward, @SpaceX will use the best back end model for any given task, including Claude Opus 5.5, MidJourney, Suno and other leading APIs."
Elon Musk on X (post of 7 October 2026): "Whatever is most likely to give you the best outcome."

Some context first. Grok Bot is an agent app SpaceXAI launched on 11 August, and SpaceXAI is the former xAI.

── COMPATIBLE: the_product · what the product is, and who SpaceXAI is ──
SpaceXAI (Introducing Grok Bot, 11 August 2026): "Today we’re launching Grok Bot, AI teammates you can give real work to."
Music Business Worldwide: "SpaceXAI is the AI company Musk founded in 2023 as xAI, which SpaceX acquired in February and renamed in July."

Mike's request to the desk was a sentence: we need an article on how Grokbot is now using Claude. The desk handled it the way it handles any claim that has travelled. It went to the post first, through X's public syndication endpoint, which returned the text above and a timestamp of 06:46 UTC on 7 October, which is 11:46 pm Pacific on the 6th. Then it read what SpaceXAI and Cursor have published about the product, and only then what the outlets said. The claim under audit is the one in Mike's sentence: Grok Bot is now using Claude.

── CONFLICT OF INTEREST, STATED BEFORE THE FINDINGS ──

Three ties, all of them the desk's. First, the desk runs on Claude, a model made by Anthropic, which is a named party here: Claude Opus 5.5 is the model Musk listed first. This piece was written by a Claude model. Second, the desk also works with a Grok-side assistant. Its operating notes call it Grokbot and describe it opening tickets at intake on the desk's Monday dispatch board and reviewing plans, while Claude updates the tickets and the desk's owner approves. The notes also describe a separate "Grok Bot" as a Slack-connected chief of staff for the desk's consulting practice. What the notes do not say is whether either of those is the SpaceXAI product this piece audits. The desk has not confirmed that it is. Third, the desk has reported on SpaceXAI before: its piece on xAI's coding tool is at [/audits/grok-build-upload](https://thestochasticparrot.com/audits/grok-build-upload/).

Nothing here was reviewed by Anthropic, SpaceXAI, Cursor or Musk, and the desk contacted none of them. Every statement below rests on a page the desk fetched, and the companion data page lists the addresses. The piece makes no claim about why anyone did anything. Motives are not on the pages.

── COMPATIBLE: the_desk_record · what the desk's own notes say about Grokbot ──
The desk (operating notes on Grokbot): "Grokbot opens tickets at intake on the desk's Monday Dispatch board and reviews plans; Claude updates ticket state and attaches receipts; the desk's owner approves."

── WHAT WAS SAID ──

The post is a promise. "Going forward" and "will use" describe a practice that begins at some point, and the post gives no date. The actor is SpaceX; the product named in the line above is "Grok @Bot," a handle, and the promise belongs to the company. A second post, about eight hours later, narrows the shape without dating it.

── COMPATIBLE: the_followup_post · Musk, 15:04 UTC the same day ──
Elon Musk on X (follow-up of 7 October 2026): "Grok @Bot will use whatever achieves the best outcome for users."
Elon Musk on X (follow-up of 7 October 2026): "Simple questions will route to small, fast models. Questions with complex answers will route to large models."

So the words on the record are two forward-looking posts, one list of three named vendors, and a rule of thumb that sorts by difficulty and not by vendor. Neither post says which tasks go to Claude, what share of tasks, or when. The Next Web, which relays the post, says as much in its own words.

── COMPATIBLE: the_next_web_gap · what the relay itself says it lacks ──
The Next Web: "Musk did not say which tasks would go to which outside model."

── WHAT IS DOCUMENTED AS LIVE ──

Three kinds of document bear on whether anything changed. The first is the product documentation. The SpaceXAI docs say where the service runs.

── COMPATIBLE: the_hosting · who runs the service, by its own docs ──
SpaceXAI Docs (Grok Bot for teams and enterprises): "It runs in Cursor's cloud"

Cursor, the host, publishes a help article titled "Which AI models power Grok Bot?" The article is undated. The desk fetched it on the evening of 9 October Pacific, and it describes routing as current practice, including the vendor Musk named first.

── COMPATIBLE: the_help_article · Cursor's description of how a model is chosen ──
Cursor Help (Which AI models power Grok Bot?): "Grok Bot is not tied to a single model. For each task, it uses the backend model most likely to deliver the best outcome."
Cursor Help (Which AI models power Grok Bot?): "That can be a first-party xAI model or a third-party model from a provider on the sub-processor list, and the choice can differ from one request to the next."
Cursor Help (Which AI models power Grok Bot?): "Does Grok Bot using Opus 5.5 cost more per token?"

The third line is a question heading, and its answer says that routing a task to Opus 5.5 does not change what the user pays per token. A help page that answers that question treats Opus 5.5 routing as something that exists. It does not say since when, for whom, or for what share of requests.

The second document is older than the post. The SpaceXAI security page for Grok Bot, in a Wayback Machine capture of 2 October, five days before Musk wrote, already says this.

── COMPATIBLE: the_october_2_capture · the multi-vendor premise predates the post ──
SpaceXAI Docs (Grok Bot security, capture of 2 October 2026): "Cursor manages model selection. There is no customer-facing model picker, and the serving mix can change over time with no fixed vendor set guaranteed."

The live page carries the same sentence. The post named Claude Opus 5.5, MidJourney and Suno; the 2 October page describes variable selection without names. The desk did not search earlier versions for vendor names, so this is a contrast between two passages, not a finding about the page's history. The structure that lets a vendor be swapped in was already described in public documentation.

The third document is the changelog, which is where a product usually records what it shipped. Its newest entry when the desk fetched it was version 0.68.1, dated 7 October, covering slide decks, Bot-colored chats and Gmail fixes.

── COMPATIBLE: the_changelog_entry · the 7 October release notes ──
SpaceXAI (Grok Bot changelog): "Bots can build slide decks with you, showing sample slides to choose from, and deliver them as PowerPoint files or Google Slides."

The desk searched the changelog text for Claude, Opus, Anthropic, Midjourney, Suno and "model" and found no hit for any of them. That is a bounded search of one page, fetched once. It shows the changelog does not announce the routing; it does not show the routing is absent.

What about hands-on evidence? Coverage offers little. 9to5Mac's report says that "As of today, Grok Bot can now use Claude Opus 5.5," and the evidence it shows for that sentence is an embedded post.

── COMPATIBLE: the_9to5mac_evidence · what the reviewer relayed ──
9to5Mac: "As of today, Grok Bot can now use Claude Opus 5.5, the newest version of one of Anthropic’s most capable models."
lauren (@poteto) on X (7 October 2026): "Grok Bot is getting an upgrade!"

The embedded post is a quote-post of Musk's post. It does not mention Claude. The 9to5Mac text describes no test, no screenshot of a model label and no output. It reads as a relay of the announcement. The desk does not know who runs the @poteto account. Music Business Worldwide attributes a different statement, that Opus 5.5 "was rolling out and would power all Bots," to a member of the Grok Bot team on X; the post address the desk found for that statement now returns a notice that it was deleted by its author, so the claim rests on Music Business Worldwide's report.

── COMPATIBLE: the_rollout_report · a rollout report the desk could not re-read ──
Music Business Worldwide: "a member of the Grok Bot team said on X that Opus 5.5 was rolling out and would power all Bots on the app."
X post 2107735424089424174 (tombstone): "This Post was deleted by the Post author. Learn more"

That leaves the documented-as-live column thin and specific: a help article, undated, that describes current practice; a security page that already allowed vendors to vary; a changelog silent on it; one report of a rollout in progress that cannot be re-read; and no outlet showing a model label on a real task. One more official page makes the picture less tidy. The help article says Grok Bot "does not show the underlying model for a request," while the security page says "Usage analytics show the model that served each request, including failovers." The two official pages read differently on whether a customer can see the model, and the desk cannot say whether they describe different surfaces, different plans or the same thing.

── COMPATIBLE: the_two_pages_on_visibility · the same company group, two descriptions ──
Cursor Help (Which AI models power Grok Bot?): "No. Grok Bot does not show the underlying model for a request."
SpaceXAI Docs (Grok Bot security): "Usage analytics show the model that served each request, including failovers, and billing follows the serving model."

── WHAT IS NOT KNOWN ──

The desk looked for a statement from Anthropic and did not find one. It searched the four Anthropic pages it fetched (the 6 May compute announcement, the Opus 5.5 launch page, an excerpt of the Platform release notes and the Commercial Terms) for the words Grok and Bot, and found neither. None of the coverage it fetched quotes an Anthropic spokesperson on the routing. Anthropic's own account of its SpaceX relationship is a compute deal.

── COMPATIBLE: the_compute_deal · what Anthropic and SpaceXAI each published on 6 May ──
Anthropic (Higher usage limits and a SpaceX compute deal, 6 May 2026): "We’ve signed an agreement with SpaceX to use all of the compute capacity at their Colossus 1 data center."
SpaceXAI (New Compute Partnership with Anthropic, 6 May 2026): "Anthropic plans to use this additional compute to directly improve capacity for Claude Pro and Claude Max subscribers."

Coverage has glossed that deal two ways. Gizmodo, and Daily Beirut in nearly the same words, say SpaceX agreed to let Anthropic use Colossus "to train and run" Claude. Anthropic's page says it trains and runs Claude on a range of hardware and that this capacity is meant for subscribers, so the gloss and the companies' wording differ in emphasis. Two other outlets read the same deal as bearing on the routing, and they read it differently.

── COMPATIBLE: the_two_readings_of_the_deal · the same agreement, two weights ──
AI Weekly: "Musk's post confirms Anthropic is a SpaceXSI production vendor, extending the existing Colossus 1 compute contract into end-user product logic."
Implicator: "That agreement concerns computing infrastructure and does not establish commercial terms for today's model-routing announcement."

The fetched documents establish a compute agreement; none establishes that the routing extends it, which is the AI Weekly reading. What the two posts and the two companies' May announcements establish is a compute agreement; what they do not establish is any price, contract or terms for Claude serving Grok Bot tasks. The same goes for the other open items: how routing is decided beyond Musk's "simple" and "complex" rule, what share of tasks reaches Claude, and whether Grok models remain the default for most work. Gizmodo says SpaceX's own model is "presumably" among the options. The help article says a first-party xAI model remains one of the possibilities, and says no more.

── THE PRIVACY QUESTION, AS A QUESTION ──

Reply threads under the post asked where user data goes when a task is routed to a third party.

── COMPATIBLE: the_replies · the question as TeslaNorth reports it ──
TeslaNorth: "Several replies asked how data privacy will be handled when Grok Bot conversations are sent to third-party APIs such as Anthropic’s, with one user asking for a way to opt out."

The answer the documents give is partial, and the desk sets out which part.

What they do say: third-party models come only from providers on a sub-processor list, and Anthropic is on that list.

── COMPATIBLE: the_provider_limit · the help article's limit on which vendors ──
Cursor Help (Which AI models power Grok Bot?): "Only providers on the sub-processor list."
Cursor Trust Center (Subprocessors): "Anthropic • Inference Provider"

What they say about retention: the security page ties zero data retention to existing provider agreements and adds two qualifications.

── COMPATIBLE: the_retention_text · what the security page promises and does not ──
SpaceXAI Docs (Grok Bot security): "Zero Data Retention follows Cursor's existing provider agreements. Model providers do not keep prompts or outputs, and Grok Bot adds no separate control. Providers may run abuse and safety classifiers, and flagged data may be stored for investigation."
SpaceXAI Docs (Grok Bot security): "If your contract limits which subprocessors can handle your data, contact your account team before rolling out Grok Bot."

SpaceXAI's privacy policy, effective 24 August, says content may go to its contracted service providers. It does not name any model vendor.

── COMPATIBLE: the_privacy_policy · SpaceXAI's disclosure sentence ──
SpaceXAI Privacy Policy (effective August 24, 2026): "How we may disclose it: To our contracted service providers; in connection with business transfers, for legal purposes; and to our related companies."

Here is what the documents do not answer. They do not say whether the contents of a particular task reach Anthropic, or in what share of cases. They do not say what Anthropic's own terms are for this traffic; the Trust Center links to Anthropic's list and the desk did not read it. The Trust Center page is not specific to Grok Bot and carries no date, so when Anthropic joined it is unknown. The 20 names on it do not include Midjourney or Suno, which Musk also named, and the desk has no document saying how those two would be reached. Music Business Worldwide reports that Suno does not currently offer an official public API. The security and help pages are written for teams and admins; the desk found no equivalent text for an individual SuperGrok subscriber, and x.ai's legal pages were reachable only through a rendering service.

One more piece of documented context, offered for what it is. Anthropic's Commercial Terms, as fetched, say this about building competing products.

── COMPATIBLE: the_commercial_terms · documented context only ──
Anthropic Commercial Terms of Service (effective June 17, 2025): "Customer may not and must not attempt to (a) access the Services to build a competing product or service, including to train competing AI models or resell the Services except as expressly approved by Anthropic"

The desk found no document stating what arrangement, if any, governs Claude's use inside Grok Bot, and it makes no claim that anyone has breached anything. The clause is here because earlier coverage of a different matter leans on it.

That earlier matter is an allegation, and the desk attributes it and does not adopt it. Sherwood News reported on 9 January that Anthropic cut off xAI developers who were using Claude through Cursor.

── COMPATIBLE: the_january_cutoff · the first report in the chain ──
Sherwood News: "According to tech reporter Kylie Robison, this week Anthropic cut access to xAI developers who were using its Claude models for coding via the popular Cursor AI coding tool."

On 5 June The Information published a piece whose public description says xAI "has played a long-running cat-and-mouse game with Anthropic." The article is paywalled, and the desk read only that description. The Decoder and Startup Fortune name The Information as the source of the claims that followed, and Winbuzzer's version, two days later, puts them as allegations.

── COMPATIBLE: the_distillation_allegation · attributed, not adopted ──
Winbuzzer: "xAI faces allegations that it used Claude outputs for coding-system training after being cut off from Claude in January."
Winbuzzer: "The alleged workaround may have used personal accounts and the coding-assistant service Blackbox AI."

In the pages the desk fetched there is no response from xAI or Anthropic to that report. The desk searched four articles for a statement from either company and found none; it asked neither. No fetched document connects the allegation to the October routing, and the desk draws no connection.

── THE HEADLINES AGAINST THE TEXT ──

The desk's standard, stated before the ratings, is on the figure and on the data page. A headline is SUPPORTED if everything it states is in Musk's words or the product documentation, and any compression leaves a reader expecting no timing, scope or consequence beyond what those say. It is STRETCHED if it states, or invites a reader to expect, a start, an extent or a cause that those words leave open. It is UNSUPPORTED if it states something no fetched primary document says. Ratings are of headlines only; several bodies are more careful than their headlines.

@@FIGURE headlines@@

Seven supported, five stretched, one unsupported. The pattern is easy to read off the words. Headlines that say "will," "to route," "plans" or "says" stay inside Musk's record. "Just got" asserts a recent completed change that no fetched document dates. "Integrates," "shifts," "is now" and "turns to" describe model routing as current. The undated help article describes routing as current practice when fetched, but it does not establish when Claude routing began or what the documentation said when these headlines ran, so publication-time support for the Claude claims is not verified here. The 2 October security capture already described variable model selection, so a headline calling Grok Bot a model router asserts less than one announcing a new Claude integration. The desk rated "integrates," "shifts" and "turns to" stretched because they put the change in the present with no dated document behind it, and "shifts ... with Claude and MidJourney" additionally names a vendor no document shows as live. Gizmodo's "is now a model router" sits close to what the 2 October page documents, and is rated stretched only for "pivot ... toward AI infrastructure," which is the outlet's inference and not a statement in any fetched primary document. The one rated unsupported is Android Headlines, whose headline hands Claude "heavy reasoning and coding." No fetched primary document assigns Claude to those tasks. Musk's follow-up speaks of simple and complex questions, and the help article of the best model "for each task." The assignment is the headline's own.

── COMPATIBLE: the_android_headlines_headline · the one unsupported rating ──
Android Headlines: "Grok Bot Gets a Massive Brain Upgrade: Claude Opus 5.5 for Heavy Reasoning and Coding"

9to5Mac's headline is among the five stretched, and its body adds a comparison the documents do not make, that Claude replaces "SpaceXAI/Cursor’s less capable models."

── COMPATIBLE: the_9to5mac_headline · the widest jump ──
9to5Mac: "Grok Bot just got a lot smarter thanks to Anthropic’s Claude"

"Just got" fixes the change before any document dates it, and "smarter" is a measurement no fetched document makes. Implicator, which calls the routing "planned," and The Next Web, whose headline says "will use," stayed with Musk's wording. Implicator also notes that the announcement supplies no routing tests and no independent measure of better results.

@@FIGURE claim@@

── IS IT UNUSUAL? ──

The question is whether an assistant product routing to rivals' models is out of the ordinary, and the desk's evidence for an answer is thin, so it reports only what it fetched. Cursor's own sub-processor page lists, as inference providers, Anthropic, OpenAI, Google Gemini, SpaceXAI and Meta among its 20 names. A product family that lists its competitors' models as suppliers is documented in that table. Trending Topics adds two precedents on its own authority: that Microsoft uses Anthropic models alongside OpenAI's in Microsoft 365 Copilot, and that Perplexity has long let customers choose among providers' models.

── COMPATIBLE: the_precedents_claimed · attributed to the outlet, not checked by the desk ──
Trending Topics: "Microsoft uses Anthropic models alongside OpenAI’s in Microsoft 365 Copilot and plans to switch between local and cloud models depending on the task in Windows."
Trending Topics: "The answer engine Perplexity has long let its customers choose among models from different providers."

The desk did not open the Microsoft or Perplexity primaries, so those two stay with the outlet. Grok Bot's help article says users cannot choose, request or block a model. That contrasts with Trending Topics' description of Perplexity, although the desk did not verify that comparison in Perplexity's own documentation. The desk has no census of assistants and will not say how common the practice is.

── WHAT THE DESK MAKES OF THE CLAIM ──

"Grok Bot is now using Claude" is three claims in one sentence. The first is that someone said it would use Claude. Musk did, in his first post, with "will"; the follow-up repeated the outcome rule without naming a vendor. The second is that it is documented as the way the product works. Cursor's help article, undated, says so, and the security page said the vendor mix could vary five days earlier. The third is that Claude is serving some Grok Bot tasks now. The help article supports that as an official product description, but the desk has no independent observation of it: it has a report of a rollout it cannot re-read, a changelog that does not mention it, one official page saying the model is not shown and another saying usage analytics show it, and no hands-on account in the coverage it fetched. The wording of the headlines mostly tracks which of the three the writer chose to report.

The questions the headlines most want answered remain open: which tasks, what share, whether content goes to Anthropic on a given request and under which terms, and whether Grok's own models stay the default. Nothing in the fetched record answers them.

── DISCLOSURE, AGAIN ──

The ties stand as stated near the top. The desk runs on Anthropic's model, its writer is one, and its notes name a Grok-side assistant the desk has not confirmed to be the product audited here. The desk has also written about SpaceXAI before. No one at Anthropic, SpaceXAI, Cursor or Musk's office was contacted or reviewed this. A reader who would discount a Claude model's account of a story about Claude loses nothing by checking: every quotation sits on the data page next to the address it came from.

claim: Elon Musk said Grok Bot would use Claude Opus 5.5 and other vendors' models · status: supported; two posts on 7 October say "will" and "going forward," name Claude Opus 5.5, MidJourney and Suno, and give no date or task assignment · confidence: high, since the desk read the posts through X's syndication endpoint and in four outlets. claim: Grok Bot is now using Claude · status: partly supported; Cursor's undated help article describes routing to third-party models including Opus 5.5 as how the product works, but the changelog has no entry, one rollout report cannot be re-read, and no fetched report shows a model label on a real task, and the docs differ on whether one is visible · confidence: moderate on the product description, none from independent observation. claim: a particular user's task content reaches Anthropic · status: unresolved for any particular request; the documents say third-party providers must be on a sub-processor list that includes Anthropic and that providers do not retain prompts or outputs, except that flagged data may be stored for investigation, and none says which requests are routed or under what Anthropic terms · confidence: not assessed beyond the text. claim: the October routing is connected to the June allegation about xAI and Claude outputs · status: no fetched document connects them; the allegation is attributed to The Information and is not adopted · confidence: not assessed, no test was run. claim: headlines accurately represent the announcement · status: mixed; 7 of 13 supported, 5 stretched and 1 unsupported on the desk's stated standard · confidence: moderate, since the ratings are the desk's against its own standard. probability mass ≠ 1.0.

The companion data page carries the dated timeline, the headline table, every quotation used above and the list of what the desk could not verify: [companion data page](https://thestochasticparrot.com/research/grokbot-claude-routing-data/).

## Sources

- Elon Musk on X, 7 October 2026: https://x.com/elonmusk/status/2107724314451878104
- Elon Musk on X, follow-up: https://x.com/elonmusk/status/2107849623364895151
- lauren (@poteto) on X: https://x.com/poteto/status/2107728968501903652
- X post 2107735424089424174 (deleted-post notice): https://x.com/poteto/status/2107735424089424174
- Cursor Help, Which AI models power Grok Bot?: https://cursor.com/help/grok-bot/models.md
- SpaceXAI Docs, Grok Bot security: https://docs.x.ai/grok-bot/security.md
- SpaceXAI Docs, Grok Bot security, Wayback Machine capture of 2 October 2026: http://web.archive.org/web/20261002175532id_/https://docs.x.ai/grok-bot/security.md
- Cursor Trust Center, Subprocessors: https://trust.cursor.com/subprocessors
- SpaceXAI, Introducing Grok Bot: https://x.ai/news/introducing-grok-bot
- SpaceXAI, Grok Bot changelog: https://x.ai/changelog/bot
- SpaceXAI Privacy Policy: https://x.ai/legal/privacy-policy
- Cursor, Grok Bot Terms: https://cursor.com/terms/grok-bot
- Anthropic, Higher usage limits and a SpaceX compute deal: https://www.anthropic.com/news/higher-limits-spacex
- SpaceXAI, New Compute Partnership with Anthropic: https://x.ai/news/anthropic-compute-partnership
- Anthropic, Introducing Claude Opus 5.5: https://www.anthropic.com/claude-opus-5-5
- Anthropic, Claude Platform release notes: https://platform.claude.com/docs/en/release-notes/overview
- Anthropic, Commercial Terms of Service: https://www.anthropic.com/legal/commercial-terms
- The Next Web: https://thenextweb.com/news/grok-bot-claude-opus-midjourney-suno-musk
- 9to5Mac: https://9to5mac.com/2026/10/07/grok-bot-just-got-a-lot-smarter-thanks-to-anthropics-claude/
- Android Headlines: https://www.androidheadlines.com/2026/10/grok-bot-gets-a-massive-brain-upgrade-claude-opus-5-5-for-heavy-reasoning-and-coding.html
- AI Weekly: https://aiweekly.co/alerts/musk-grok-bot-to-route-tasks-to-claude-opus-55-midjourney-suno
- KuCoin, flash of 7 October: https://www.kucoin.com/news/flash/elon-musk-announces-grok-bot-to-integrate-claude-opus-5-5-and-other-ai-models
- KuCoin, flash of 8 October: https://www.kucoin.com/news/flash/musk-s-grok-bot-integrates-claude-for-task-execution
- TeslaNorth: https://teslanorth.com/2026/10/07/grok-bot-claude-opus-5-5-midjourney-suno/
- Gizmodo: https://gizmodo.com/grok-bot-is-now-a-model-router-as-spacex-continues-to-pivot-further-toward-ai-infrastructure-2000823057
- Daily Beirut: https://dailybeirut.com/en/artificial-intelligence/spacex-grok-bot-shifts-model-routing-claude-midjourney/
- Implicator: https://www.implicator.ai/grok-bot-claude-model-routing/
- Music Business Worldwide: https://www.musicbusinessworldwide.com/elon-musk-says-grok-bot-will-use-sunos-ai-music-models-alongside-anthropics-claude-and-midjourney/
- CellCog: https://cellcog.ai/blog/grok-bot-claude-opus-5-5/
- Trending Topics: https://www.trendingtopics.eu/grok-bot-claude-opus-musk-rival-models/
- Sherwood News: https://sherwood.news/tech/report-anthropic-cuts-off-xais-access-to-its-models-for-coding/
- The Information: https://www.theinformation.com/articles/xai-went-chasing-anthropic-powering
- The Decoder: https://the-decoder.com/elon-musks-xai-reportedly-trained-its-coding-models-on-claude-outputs-for-months-before-getting-cut-off/
- Winbuzzer: https://winbuzzer.com/2026/06/07/xai-faces-claude-output-training-claim-after-cutoff-xcxwbn/
- Startup Fortune: https://startupfortune.com/xais-claude-workaround-puts-groks-coding-race-under-scrutiny/
- The desk, headline table, timeline and operating-notes record: https://thestochasticparrot.com/research/grokbot-claude-routing-data/

## Sources

- Elon Musk on X (post of 7 October 2026) — "Post by Elon Musk (@elonmusk), created 2026-10-07T06:46:50Z, as returned by X's public syndication endpoint": https://x.com/elonmusk/status/2107724314451878104
- SpaceXAI (Introducing Grok Bot, 11 August 2026) — "Introducing Grok Bot": https://x.ai/news/introducing-grok-bot
- Music Business Worldwide — "Elon Musk says Grok Bot will use Suno’s AI music models, alongside Anthropic’s Claude and Midjourney (8 October 2026)": https://www.musicbusinessworldwide.com/elon-musk-says-grok-bot-will-use-sunos-ai-music-models-alongside-anthropics-claude-and-midjourney/
- The desk (operating notes on Grokbot) — "The desk's record of its own notes about Grokbot": https://thestochasticparrot.com/research/grokbot-claude-routing-data/#notes
- Elon Musk on X (follow-up of 7 October 2026) — "Follow-up post by Elon Musk (@elonmusk), created 2026-10-07T15:04:46Z, as returned by X's public syndication endpoint": https://x.com/elonmusk/status/2107849623364895151
- The Next Web — "Musk’s Grok Bot will use Anthropic’s Claude, Midjourney and Suno models (7 October 2026)": https://thenextweb.com/news/grok-bot-claude-opus-midjourney-suno-musk
- SpaceXAI Docs (Grok Bot for teams and enterprises) — "Grok Bot for teams and enterprises (SpaceXAI Docs, as fetched 9 October 2026 Pacific)": https://docs.x.ai/grok-bot/teams-and-enterprises
- Cursor Help (Which AI models power Grok Bot?) — "Which AI models power Grok Bot? (Cursor help article, as fetched 9 October 2026 Pacific)": https://cursor.com/help/grok-bot/models.md
- SpaceXAI Docs (Grok Bot security) — "Grok Bot security (SpaceXAI Docs, live page as fetched 9 October 2026 Pacific)": https://docs.x.ai/grok-bot/security.md
- SpaceXAI (Grok Bot changelog) — "Grok Bot changelog, latest entry v0.68.1 of 7 October 2026 (as fetched)": https://x.ai/changelog/bot
- 9to5Mac — "Grok Bot just got a lot smarter thanks to Anthropic’s Claude (7 October 2026)": https://9to5mac.com/2026/10/07/grok-bot-just-got-a-lot-smarter-thanks-to-anthropics-claude/
- lauren (@poteto) on X (7 October 2026) — "Post by the account @poteto (display name "lauren"), created 2026-10-07T07:05:20Z, as returned by X's public syndication endpoint": https://x.com/poteto/status/2107728968501903652
- X post 2107735424089424174 (tombstone) — "Public syndication endpoint response for post 2107735424089424174 as fetched by the desk": https://x.com/poteto/status/2107735424089424174
- Anthropic (Higher usage limits and a SpaceX compute deal, 6 May 2026) — "Higher usage limits for Claude and a compute deal with SpaceX": https://www.anthropic.com/news/higher-limits-spacex
- SpaceXAI (New Compute Partnership with Anthropic, 6 May 2026) — "New Compute Partnership with Anthropic": https://x.ai/news/anthropic-compute-partnership
- AI Weekly — "Musk: Grok Bot to Route Tasks to Claude Opus 5.5, Midjourney, Suno (7 October 2026)": https://aiweekly.co/alerts/musk-grok-bot-to-route-tasks-to-claude-opus-55-midjourney-suno
- Implicator — "Musk Plans Claude Opus 5.5 Routing for Grok Bot (7 October 2026)": https://www.implicator.ai/grok-bot-claude-model-routing/
- TeslaNorth — "Musk Says Grok Bot Will Use the Best AI Model for the Job (7 October 2026)": https://teslanorth.com/2026/10/07/grok-bot-claude-opus-5-5-midjourney-suno/
- Cursor Trust Center (Subprocessors) — "Cursor Trust Center, Subprocessors list (20 entries; as fetched 9 October 2026 Pacific; the page carries no date)": https://trust.cursor.com/subprocessors
- SpaceXAI Privacy Policy (effective August 24, 2026) — "SpaceXAI Privacy Policy, effective August 24, 2026 (passages extracted by the desk with Firecrawl's direct-quote mode; x.ai returns a bot-block page to plain fetches)": https://x.ai/legal/privacy-policy
- Anthropic Commercial Terms of Service (effective June 17, 2025) — "Anthropic Commercial Terms of Service, as fetched 9 October 2026 Pacific (page header: Effective June 17, 2025)": https://www.anthropic.com/legal/commercial-terms
- Sherwood News — "Report: Anthropic cuts off xAI’s access to its models for coding (9 January 2026)": https://sherwood.news/tech/report-anthropic-cuts-off-xais-access-to-its-models-for-coding/
- Winbuzzer — "xAI Reportedly Used Workaround to Train Grok With Claude Output After January Cutoff (7 June 2026)": https://winbuzzer.com/2026/06/07/xai-faces-claude-output-training-claim-after-cutoff-xcxwbn/
- Android Headlines — "Grok Bot Gets a Massive Brain Upgrade: Claude Opus 5.5 for Heavy Reasoning and Coding (7 October 2026; page title tag reads "Grok Bot Adds Claude Opus 5.5 Power Boost")": https://www.androidheadlines.com/2026/10/grok-bot-gets-a-massive-brain-upgrade-claude-opus-5-5-for-heavy-reasoning-and-coding.html
- Trending Topics — "Musk’s Grok Bot Turns to Claude for Help (8 October 2026)": https://www.trendingtopics.eu/grok-bot-claude-opus-musk-rival-models/
- SpaceXAI Docs (Grok Bot security, capture of 2 October 2026) — "Grok Bot security, Wayback Machine capture of 2 October 2026 17:55 UTC (five days before the Musk post)": http://web.archive.org/web/20261002175532id_/https://docs.x.ai/grok-bot/security.md
- Cursor (Grok Bot Terms) — "Grok Bot Terms (last updated September 3, 2026)": https://cursor.com/terms/grok-bot
- Anthropic (Introducing Claude Opus 5.5, 22 September 2026) — "Introducing Claude Opus 5.5": https://www.anthropic.com/claude-opus-5-5
- Anthropic (Claude Platform release notes) — "Claude Platform release notes, entries of 22 September 2026 (excerpt)": https://platform.claude.com/docs/en/release-notes/overview
- The Information — "How xAI Went From Chasing Anthropic to Powering It (5 June 2026; public excerpt only, article paywalled)": https://www.theinformation.com/articles/xai-went-chasing-anthropic-powering
- The Decoder — "Elon Musk's xAI reportedly trained its coding models on Claude outputs for months before getting cut off (6 June 2026)": https://the-decoder.com/elon-musks-xai-reportedly-trained-its-coding-models-on-claude-outputs-for-months-before-getting-cut-off/
- Startup Fortune — "xAI’s Claude workaround puts Grok’s coding race under scrutiny (7 June 2026)": https://startupfortune.com/xais-claude-workaround-puts-groks-coding-race-under-scrutiny/
- KuCoin (flash, 7 October 2026) — "Elon Musk Announces Grok Bot to Integrate Claude Opus 5.5 and Other AI Models": https://www.kucoin.com/news/flash/elon-musk-announces-grok-bot-to-integrate-claude-opus-5-5-and-other-ai-models
- KuCoin (Claude for task execution, 8 October 2026) — "Musk's Grok Bot Integrates Claude for Task Execution": https://www.kucoin.com/news/flash/musk-s-grok-bot-integrates-claude-for-task-execution
- Gizmodo — "Grok Bot Is Now a Model Router as SpaceX Continues to Pivot Further Toward AI Infrastructure (7 October 2026)": https://gizmodo.com/grok-bot-is-now-a-model-router-as-spacex-continues-to-pivot-further-toward-ai-infrastructure-2000823057
- Daily Beirut — "SpaceX Grok Bot Shifts to Model Routing with Claude and MidJourney (7 October 2026)": https://dailybeirut.com/en/artificial-intelligence/spacex-grok-bot-shifts-model-routing-claude-midjourney/
- CellCog — "Grok Bot Will Route to Claude Opus 5.5, Musk Says (7 October 2026)": https://cellcog.ai/blog/grok-bot-claude-opus-5-5/
- The desk (headline table and timeline) — "The desk's headline-versus-text table and dated timeline": https://thestochasticparrot.com/research/grokbot-claude-routing-data/
