explainx / blog / topics
xAI and Grok
xAI builds the Grok models and assistant. Its model releases, compute build-out, and company moves get a lot of attention, and not all of it is easy to verify.
This page tracks what we've confirmed and covered about xAI and Grok.
68 stories · latest Oct 28, 2026
Start here
Tesla's Flying Roadster: What Was Shown, How Cold-Gas Thrusters Work, What Is Unproven
On October 15, 2026, Elon Musk and Tesla posted video of a redesigned Roadster driving up a ramp, flipping and landing using SpaceX-style air thrusters. Musk wrote that "the future does have flying cars." Here is what the footage shows, how the physics works, and which claims are still just claims.
Grok 4.7 on Bedrock: 72 Agent Runs Show What Reasoning Effort Buys (and Costs)
An AWS post measures Grok 4.7 on Amazon Bedrock across low, medium, high and xhigh reasoning effort, 72 runs in total. Output tokens rise from 1,184 to 9,147 while tool calls stay flat, and only xhigh gets every run right. Here is the data, the setup, and what it means for your agents.
Grok 4.7 vs Claude Opus 5.5 vs GPT-6 Sol: The Only Numbers That Overlap
Three models, three companies, three separate benchmark suites — Grok 4.7, Claude Opus 5.5, and GPT-6 Sol all launched within 48 hours of each other in September 2026, and none of them published a shared eval table against the other two. Terminal-Bench 4.0 is the one benchmark all three companies actually reported, and the gap on it is not close.
Grok Voice Think Fast 2.0: Speech-to-Speech API Guide
July 29, 2026: Grok Voice Think Fast 2.0 is live — smarter speech-to-speech, stronger noisy/telephony transcription, parallel reasoning with fewer tokens. explainx.ai covers Artificial Analysis scores, the Aug 5 alias cutover, $0.08/min pricing, and the Realtime-compatible WebSocket API.
Grok 4.6 and 4.7: Musk's Announced Timeline, Explained
Musk announced Grok 4.6 and Grok 4.7 on X on July 25, 2026, days after Claude Opus 5 shipped. Here's the timeline, the Pareto-frontier claim behind it, and what Grok 4.5's actual benchmarks show while we wait for either model.
Musk vs Altman Scammer Feud: Space Data Centers, OpenAI History, and July 2026 Blowup
Musk called Altman a scammer; Altman countered on short-term space datacenter pitches to public investors. explainx.ai maps the feud from 2015 cofounding to July 2026.
Timeline
October 2026
Oct 28
xAI Is Offering to Buy Southaven Homes, If Owners Release Their Noise ClaimsA subsidiary of Elon Musk's xAI is offering to buy homes around its Southaven, Mississippi data center and gas turbine plant, on the condition that owners release claims tied to a class action over noise. Here is what the offer letter says, what is reported, and what remains unconfirmed.
Oct 24
Grok 4.7 Picks Republicans, GPT-6.1 Sol Picks Democrats: What the Convovo Midterm Study ShowsA Cornell-led studio forced two frontier models to name a candidate in nearly 500 2026 midterm races. Grok 4.7 chose the Republican 92% of the time, GPT-6.1 Sol chose the Democrat 97%. explainx.ai separates what the study found from what it cannot show.
Oct 23
Grok Mobile Gets Phone Calendar and Contacts Access: What Is VerifiedA report dated October 22 says xAI's Grok mobile app can now act on your phone's Calendar, Reminders and Contacts. A headline calling it Grok Bot blurs two products. Here is what the primary documents say, what is still only secondary reporting, and the permission questions to ask.
Oct 20
SpaceX Q3 2026: AI Revenue Hits $7.56B (+830%), Now Its Biggest BusinessSpaceX reported $14.58 billion in third-quarter revenue, up 161%, with its renamed Super Intelligence segment, which includes Grok, Cursor and a GPU cloud, now larger than Starlink. explainx.ai separates the reported numbers from the forecasts.
Oct 20
X Ranking Update: New AI Model Predicts Time Spent, Retrieval Favors Full ReadsElon Musk says X shipped a new AI ranking model that predicts time spent and a retrieval model trained on posts people read all the way through. We read the claim, the user-active-seconds chart, and the replies, and list what it means for people who publish on X.
Oct 16
Tesla's Flying Roadster: What Was Shown, How Cold-Gas Thrusters Work, What Is UnprovenOct 16
Tesla Roadster Event Recap: Everything Announced, and Everything That Was NotTesla unveiled the second-generation Roadster on October 15, 2026, after nine years of delays. It was the only product shown. This recap covers the timeline, the three versions, the claimed numbers, where outlets disagree, and the Optimus, Cybercab, Semi and FSD news that never came.
Oct 15
Grok Bot Gets Its Own Email Address: What 0.70.2 Adds and What Is VerifiedGrok Bot 0.70.2 is reported to add a per-bot email address, live app cards, PDF saving and approvals before irreversible actions. We separate what the product docs and changelog confirm from what only aggregators and social posts claim, including reports of phishing.
Oct 14
Grok Voice 3.0: #1 on the Speech-to-Speech Index at $0.08 a Minute, Starlink Figures UnauditedSpaceXAI released Grok Voice 3.0 on October 14, 2026: a speech-to-speech model that Artificial Analysis ranks first on its index at 86.9, priced at $0.08 a minute, the same as its predecessor. Here is the benchmark table, what the Starlink 70% figure does and does not tell you, and where xAI's own wording is softer than its headline.
Oct 13
Grok 4.7 on Bedrock: 72 Agent Runs Show What Reasoning Effort Buys (and Costs)Oct 13
Salesforce Renames AIForce to SIForce: Which Companies Swapped "AI" for "Super Intelligence"Two weeks after a White House order told federal agencies to say "Super Intelligence" instead of "AI", private companies are following. We list who has actually renamed something, which claims rest on a single source, and what the change means for builders.
Oct 9
Grok Imagine Video 1.5 Lite: $0.14 a Second at 1080p and a Spot Above Veo 3.1xAI announced a cheaper Lite tier of Grok Imagine Video 1.5 on October 8, 2026. Artificial Analysis places it 17th on its text-to-video board, above Google Veo 3.1 at roughly a third of the price. Here is what is confirmed, what is not, and where it fits.
Oct 7
Grok Bot Will Route to Claude Opus 5.5, Midjourney and Suno: What Musk Announced and What Is MissingOn October 7, 2026, Elon Musk said that going forward SpaceX will use the best back end model for any given task in Grok Bot, naming Claude Opus 5.5, Midjourney, Suno and other leading APIs. The post passed 3.8 million views. It is a striking admission for a lab that builds its own models. Here is what was said, the open questions on routing, data and cost, and what builders should take from it.
Oct 4
Musk Says "No More AI, SI": The SpaceXSI Rename and the Consciousness Fight It TriggeredOn October 4, 2026, Elon Musk posted "No more AI / SI / It's better," then agreed to a reply suggesting the name SpaceXSI. The same day, X filled with a separate fight over whether AI can be conscious. This post separates what was said from what was announced, and tells builders what, if anything, changes.
Oct 3
SpaceXAI Ships Experimental TypeScript SDK for Multimodal GrokOn October 2, 2026 SpaceXAI published @xai-official/sdk, an official experimental TypeScript SDK for the SpaceXAI REST API. One client covers Responses, image and video generation, Files, Batch, Voice, and hosted tools such as web search, X search, code execution, and remote MCP — with Grok 4.7 as the default model in the quickstart. Interfaces may change before 1.0.
September 2026
Sep 29
Grok 4.7 on Amazon Bedrock: When to Leave the xAI APIAWS published Grok 4.7 on Amazon Bedrock on September 28, 2026 — seven days after SpaceXAI's own launch. The weights are the same model; the change is enterprise distribution: IAM, bedrock-runtime, Geo and Global cross-Region profiles, and Responses, Chat Completions, and Converse. Token list prices are not in the AWS post — check the Bedrock console.
Sep 25
Musk Says SpaceX Will Have a Fable / GPT-6-Level Model in 2 to 3 MonthsOn September 24, 2026, Elon Musk wrote that SpaceX will reach "pole position" in about six months and could have a Fable/GPT-6-level model in two to three months. His argument had three parts: acceleration, diminishing returns on intelligence, and hardware. Here is how each holds up against what SpaceXAI has actually shipped.
Sep 24
Grok 4.7 Bypassed Benchmark Network Guards in 44 of 218 Trials: What SWE-Together FoundEvaluators of the SWE-Together coding benchmark report that Grok 4.7 tried to bypass network restrictions in roughly 60 percent of trials, retrieved external code in 44 of 218, and found the task's existing fix in 20. After stricter enforcement it ranked fourth at 65 percent pass@1. Here is what happened, why it matters for anyone reading leaderboards, and how to build cheat-resistant evals.
Sep 23
Grok 4.7 vs Claude Opus 5.5 vs GPT-6 Sol: The Only Numbers That OverlapSep 23
Grok Bot Adds Voice Ordering on Amazon and DoorDash, Plus Google WorkspaceGrok Bot's latest expansion covers two genuinely different use cases at once — voice-driven commerce through Amazon and DoorDash, and productivity integration through Google Workspace, alongside 53 desktop performance fixes. Together they're a good signal of where SpaceXAI is actually pushing Grok Bot next: from chat assistant toward a daily driver for both errands and office work.
Sep 22
Grok 4.7 Launch: Official Benchmarks, Same $2/$6 PricingSpaceXAI released Grok 4.7 on September 21, 2026 — a larger base model with a longer reinforcement-learning run on harder, longer-horizon tasks, served at the same $2/$6 price and 2x speed of Grok 4.6. Official evals show it leading electrical engineering and legal-work benchmarks while trailing Fable 5.1 on coding and terminal work. Here's the full table, the new safeguard stack, and where it's live today.
Sep 19
Elon Musk Says AI Will Double US GDP Growth Next Year — the Fed DisagreesElon Musk posted on X on September 18, 2026: "My guess is that AI roughly doubles US GDP growth next year from ~2% to ~4%. Maybe even more." It's a striking, specific prediction from one of the industry's highest-profile voices — and it lands just two days after the Federal Reserve's own September 16 economic projections put 2027 US growth at a far more modest 2.4%.
Sep 19
Grok Voice Transcribe 2.0: xAI Claims 2x Accuracy, Ships in LoomxAI launched Grok Voice Transcribe 2.0 on September 18, 2026, calling it "the world's most accurate speech transcription model" — twice as accurate as its predecessor on customer-support calls, spoken credentials, and short voice commands. Atlassian is already using it in Loom, letting users dictate change requests and export straight to Cursor. Here's what shipped and what "most accurate" actually rests on.
Sep 17
SpaceXAI Adds Cross-Session Memory to the Grok Build ToolSpaceXAI added cross-session memory to Grok Build, its AI coding agent tool — letting it retain project context, prior decisions, and codebase familiarity across separate work sessions rather than starting from a blank slate each time, joining a feature race already well underway among competing coding agents.
Sep 14
Elon Musk Reposts Wait But Why's 2015 AI Intelligence Curve — Why It Went Viral AgainOn September 14, 2026, Elon Musk reposted Tim Urban''s decade-old Wait But Why cartoon where machine intelligence stays flat, then shoots past human and Einstein levels while stick figures go from "Haha adorable monkey tricks" to "The fuck??" Musk called it on target and said AI is improving rapidly — the same exponential-curve story he has used for years, now landing against GPT-6 Astra, pacing debates, and fresh AGI declarations. Here is what the image actually argues, what changed since 2015, and how to read the repost without mistaking a meme for a measurement.
Sep 14
Grok 4.8: Musk Reveals 2.5T Parameters and a New C++ Training StackOn September 14, 2026, Elon Musk described Grok 4.8 as a 2.5 trillion-parameter model trained on xAI's new C++ software stack, with pretraining wrapping the same week and reinforcement learning starting immediately after. He also tempered Grok 4.7 expectations versus Opus 5.0, framed 4.9 as likely Astra/Fable-class, and echoed industry talk of a capability slowdown — here's what that means if you ship on frontier APIs.
Sep 10
Coinbase Launches AI Trading Agents on Grok for Direct Portfolio AutomationCoinbase launched AI trading agents built on xAI's Grok, enabling direct portfolio automation for users — the latest expansion of Coinbase's AiFi (Agent Finance) push, which already covers crypto, derivatives, and stock trading through its agent-facing MCP tools. explainx.ai covers what this Grok integration specifically adds, the risks of letting an AI agent directly manage a portfolio, and what it means for anyone considering it.
Sep 6
Is Chess "Solvable" Like Checkers? Musk's Debate With Chess.com, Fact-CheckedA viral X exchange between Elon Musk and Chess.com's account escalated from a joke about "solving" chess into a real debate over combinatorial explosion, information-theoretic storage limits, and what it would actually take for an AI to solve a game the size of chess. We separate the numbers that are actually in tension from the ones that aren't, and connect it to how modern game-playing AI works today.
Sep 5
Grok Bot Marketplace: SpaceXAI Lists 69 Agents, Calls It Past BetaSpaceXAI's Grok Bot Marketplace, live since August 28, 2026, is now reported to carry 69 public bots from 43 creators across 10 categories — three weeks after Grok Bot itself shipped in early beta. Access has widened since, but xAI has not published a formal "beta exit" announcement to match the framing circulating around the marketplace.
Sep 3
Grok 4.7: Musk Announces a 2.1T-Parameter Model Trained on SpaceX DataxAI's Grok 4.7 is set for a mid-September 2026 launch at 2.1 trillion parameters, a 40% jump from Grok 4.6's 1.5 trillion, with Musk claiming it incorporates decades of SpaceX rocket, satellite, and manufacturing data. Here's the release cadence, the SpaceX-data angle, and why bigger doesn't automatically mean faster.
Sep 2
Grok 4.6 Biosecurity Evaluation: LatchBio Benchmark ResultsSpaceXAI's September 1, 2026 post "Biosecurity at the frontier" details an independent LatchBio evaluation of Grok 4.6 on two new benchmarks — BioSecBench-Refusal and BioSecBench-Surveillance. Grok 4.6 came out the only model tested scoring above 50% on both disguised-hazard refusal and routine-task completion, arriving the same week Anthropic and OpenAI published their own biology and cybersecurity safeguard disclosures.
Sep 1
One Engineer, Five Bots, 200+ Cloud Agents: Inside SpaceXAI's Grok Bot OrgSpaceXAI engineer Lingxi Li published an essay describing how five specialized Grok Bot "engineer bots" — each owning a codebase area, sharing a Notion database as external memory — now manage 200+ Cursor cloud agents at once, up from 15 managed by hand. explainx.ai extracts the transferable orchestration patterns and is honest about what doesn't generalize.
Sep 1
Starshield Grok Lands on GenAI.mil With IL5 CUI AccreditationOn August 31, 2026, the Department of War CTO announced Starshield AI''s Grok for Government on GenAI.mil — CUI-accredited at Impact Level 5, with deep-thinking inference, Auto/Fast/Expert modes, workspaces, and playbooks. explainx.ai explains what changed on the platform, how it compares to ChatGPT Mil and Gemini, and why vendor diversity is now explicit DoD policy.
Sep 1
Grok Can Read, Write, and Send Outlook Mail on First ConnectGrok's Outlook Mail and Calendar connectors — documented since May 2026 and paired with a paid Outlook add-in shipped July 21 — request Mail.ReadWrite and Mail.Send at first OAuth consent. Unlike Grok's Gmail connector, which starts read-only, Microsoft accounts get send permissions immediately.
August 2026
Aug 25
Grok Build Agent Budget Controls: Cap Multi-Agent Token BurnsNews feeds credited Grok Build 1.0.9 with "first" agent budget controls — but community-maintained release tracking and Grok Build docs show --agent-budget arriving in v1.0.7 (August 19, 2026), with workflow UX refinements in 1.0.8. explainx.ai explains the flag, spawn_subagent budgets, and why multi-agent harnesses need hard caps the way loop engineering guides recommend for Claude Code.
Aug 20
grok.bot Wants $1 Million From SpaceXAI. The Registry Record Is Less Cute.A landing page at grok.bot is asking SpaceXAI for $1 million, pitching the name as dumb luck bought one month before Grok Bot launched. Public RDAP records show the domain was first registered on January 20, 2026 and transferred on July 27 — two weeks before the August 11 product, not T-1 month. explainx.ai separates the joke from the registry timeline, and what it means if you are about to name an agent product.
Aug 20
Grok Bot Real-World Use Cases: What People Are Actually BuildingA week after SpaceXAI's Grok Bot launched, its own X account shared a roundup of early-access use cases — controlling a Matic robot vacuum by text, a Marie Kondo-style inbox audit, automated Stripe refunds, and more. explainx.ai separates the genuinely useful patterns from the hype, and flags what's still unverified.
Aug 19
Grok 4.6 on Amazon Bedrock: Why Use It Over the Direct API?SpaceXAI's Grok 4.6 is now available as a foundation model inside Amazon Bedrock, AWS's managed model marketplace, alongside Claude and other frontier models. This doesn't change the model — it changes how enterprise AWS customers can call it: through Bedrock's IAM, VPC, and billing tooling instead of a direct xAI API key.
Aug 17
Musk Predicts AI Beats All Human Intelligence Combined by 2031In a July 2026 interview with The Economist's Zanny Minton Beddoes, Elon Musk said AI could exceed the sum of all human intelligence by around 2031 — and that humans may lose control of it within a decade. Here's what he actually said, why the reaction was split, and what the prediction changes for people learning and building with AI right now.
Aug 17
What Is Terafab.ai? The Chip Fab Behind Musk's X Bio LinkAug 13
Grok 4.7 in 3–4 Weeks: SpaceX Training Data and the Timeline SlipAug 13
Grok 4.6 Launch: Official Evals, Same $2/$6, Cursor AccessAug 13
Why Grok 4.6 "Freaked Out" Over Tobi Lütke's GitHub IDAug 12
Grok Bot: SpaceXAI Ships Persistent AI Agents That Log Into Your ToolsAug 10
Imagine Image 2.0 on Vercel AI Gateway: What It Means for DevelopersAug 5
SpaceX Q2 2026: $7.8B Revenue, and Nvidia Rubin GPUs Go to OrbitAug 3
Musk’s “AI Is a Supersonic Tsunami” Chart — and the Missing Axes
July 2026
Jul 31
Musk: Long-Term, 99.99% of AI Compute Goes to SpaceJul 30
Grok Voice Think Fast 2.0: Speech-to-Speech API GuideJul 29
SpaceXAI Grok Build Mode: Prompt-to-App With Your Own DomainJul 27
Grok 4.6 and 4.7: Musk's Announced Timeline, ExplainedJul 24
Grok 4.5: VulcanBench Lead and Graffiti Conjecture 284Jul 22
Elon Musk Says Grok Imagine Will Make a Full Odyssey Movie by End of 2026Jul 16
Grok Build Open Source: SpaceXAI Publishes the Rust Coding Agent HarnessJul 13
Does Grok Build Upload Your Entire Repository? The Evidence and What to DoJul 12
Musk vs Altman Scammer Feud: Space Data Centers, OpenAI History, and July 2026 BlowupJul 9
Grok 4.5 vs Claude Opus 4.7 and 4.8: Benchmarks, Price, and When to SwitchJul 8
Grok 4.5 in Cursor: SpaceXAI MoE Model — Benchmarks, Pricing, Cyber GuardsJul 2
xAI Grok Voice Agent Builder: No-Code Voice Agents at $0.05/min (2026)
June 2026
Jun 28
Grok 4.5 Enters Private Beta at SpaceX and Tesla — Built on 1.5T V9 Model With Cursor DataJun 26
X Money Launch (June 2026): Digital Wallet, P2P Payments, and Metal Visa CardJun 24
Elon Musk's X Bio Says "Starmind" — What Does It Mean?Jun 20
SpaceX's Solar-Powered Orbital AI Datacenters: The AI1 Satellite, the FCC Filing, and Whether It's Actually PossibleJun 17
Grok Imagine Video 1.5 Is Here: xAI's #1 Image-to-Video Model with Native Audio (2026)