xAI has officially released Grok 4.6, the next iteration of its flagship large language model, positioning it as a substantial jump over Grok 4.5 at the same API price. Elon Musk called the release a 'banger' on X, citing a 1753 ELO score and pricing that undercuts other frontier models by roughly half. The model is live today across Grok Build, Cursor, Grok Bot, and the xAI API, with a first-week promotion doubling included usage in Cursor and Grok Build.
What xAI Is Claiming
The official announcement thread from the xAI account frames Grok 4.6 as a step-change on three axes: intelligence, speed, and cost. According to xAI, the model 'delivers frontier intelligence and is a significant improvement over Grok 4.5 at the same price,' and 'can handle much more challenging tasks' than its predecessor. Musk followed up with a benchmark claim — a 1753 ELO score — and the marketing pitch of 'amazing bang for buck.'
According to independent reporting from OrcaRouter and kie.ai, Grok 4.6 is a 1.5-trillion-parameter model built on the same V9 foundation as Grok 4.5, but with a longer supplemental training run that layered in improved supervised fine-tuning and reinforcement learning on curated data. In other words: same base architecture, materially more training work on top.
Key Figures
| Metric | Grok 4.6 | Notes |
|---|---|---|
| API input pricing | $2 / M tokens | xAI states this is half the price of other frontier models |
| API output pricing | $6 / M tokens | Faster variant available at 2x this rate, per OrcaRouter |
| ELO score (xAI-cited) | 1753 | Cited by Musk; not yet independently verified on public arenas |
| Parameter count | 1.5 trillion | Same V9 base as Grok 4.5, per OrcaRouter reporting |
| Launch date | Aug 7, 2026 (rollout); public announcement Aug 12 | Hits target Musk set publicly on July 28 |
| Promo | 2x included usage | First week only, Cursor + Grok Build |
The Pricing Story Is the Real Headline
The ELO number will get the retweets, but the pricing is arguably the bigger competitive move. At $2 per million input tokens and $6 per million output tokens, Grok 4.6 sits well below the sticker price of comparable frontier tier offerings from OpenAI and Anthropic. xAI's messaging is explicit: 'half the price of other frontier models.' That framing is aimed squarely at developers and agent builders whose bills scale with token volume — the exact cohort xAI needs to win to catch up on API market share.
Pairing that pricing with a first-week 2x usage promotion inside Cursor — one of the most popular AI-native code editors — is a distribution play. Cursor's developer audience is where competitive AI model preferences get decided in real workflows, not on Twitter benchmarks. Doubling usage during launch week means Cursor users will actively test Grok 4.6 against whatever model they're currently defaulting to, at essentially no marginal cost.
The 1753 ELO Claim, With Caveats
Musk's 1753 ELO figure is worth flagging carefully. According to independent tracking on prediction market platforms like PredictionHunt, as of August 11 — one day before this announcement — public prediction markets were still re-evaluating the likelihood of Grok 4.6 hitting high ELO scores on platforms like LMSYS Chatbot Arena, and independent benchmarks had not yet published. The 1753 number is xAI-cited and appears to reflect internal or early third-party arena scoring. Public leaderboards typically take days to weeks to stabilize as vote volume accumulates.
Separately, xAI states Grok 4.6 matches GPT-5.6 Sol on the Artificial Analysis Intelligence Index with a composite score of 61, according to reporting from OfficeChai and kie.ai. That's a more structured benchmark and, if it holds up in independent testing, would put Grok 4.6 in the same tier as the current frontier — at half the price.
Where You Can Use It Today
Grok 4.6 is live as of the announcement across four primary surfaces: the xAI API, Grok Build (xAI's developer platform), Cursor (the AI coding editor), and Grok Bot on X. Third-party access is also live via OpenRouter, Vercel, and Cloudflare, according to xAI's launch materials — meaning developers who route requests through model marketplaces can switch to Grok 4.6 without changing providers.
Existing Grok subscribers on X should see the model rotate in automatically. Cursor users need to select Grok 4.6 from the model dropdown to trigger the 2x usage promo. API developers can point requests at the new model identifier immediately.
What This Means for the Tesla-Adjacent AI Story
For Tesla owners tracking xAI's arc, Grok 4.6 matters because xAI is now the AI supplier being wired into Tesla vehicles for in-car voice interaction and, longer term, into Optimus. A cheaper, faster, higher-scoring Grok makes the economics of shipping AI features across millions of Tesla endpoints materially better. Every cent of token cost gets multiplied by fleet scale.
The cadence xAI is signaling is aggressive. According to OrcaRouter, Grok 4.7 — a larger 2.1-trillion-parameter model — is expected within weeks of the 4.6 release, and Grok 5 is targeted before the end of 2026. If those dates hold, xAI is pushing a release tempo that few frontier labs have sustained.
What to Watch Next
- Independent ELO confirmation. LMSYS Chatbot Arena and Artificial Analysis will post community-verified scores within days. Whether the 1753 figure holds up publicly is the first credibility test.
- Cursor adoption data. If Grok 4.6 becomes a default in a meaningful share of Cursor sessions after the 2x promo ends, that's the strongest signal the pricing pitch is working.
- Grok 4.7 timing. A larger 2.1T-parameter model was reported as imminent. A shipping date confirms the aggressive release cadence.
- Tesla in-car integration. Watch for a Tesla OTA that references an updated Grok voice assistant model — the natural downstream beneficiary of a cheaper, faster frontier tier.
The next 72 hours of independent benchmark data will decide whether Grok 4.6 is genuinely competitive at the top of the frontier stack, or whether the ELO claim outpaces the reality. Either way, the pricing move is real, it's live today, and it puts direct cost pressure on every other frontier lab's API business.
Related Gear
Gear up your Tesla with tested, custom-fit BASENOR accessories — shop Tesla accessories →
Sources & reporting notes
The links below identify the material source records used for this report.
- @SpaceXAI on X (2026-08-12T15:32:11.000Z) — Direct source
- @SpaceXAI on X (2026-08-12T15:32:11.000Z) — Direct source
- @SpaceXAI on X (2026-08-12T15:32:12.000Z) — Direct source
- @michaelnicollsx on X (2026-08-12T15:36:34.000Z) — Direct source
- @elonmusk on X (2026-08-12T15:41:00.000Z) — Direct source
- @elonmusk on X (2026-08-12T15:42:25.000Z) — Direct source
- @elonmusk on X (2026-08-12T15:46:31.000Z) — 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.
