SpaceXAI releases Grok 4.7, which it says is better at verifying its own work and managing longer context, available for $2/1M input and $6/1M output tokens
SpaceXAI's most powerful model for coding and knowledge work. Twice as fast, at half the price of comparable models.
xAI
Context & Ripple Effects
Grok’s release cadence has moved from the multi-agent Grok 4 Heavy tier in 2025 to a Cursor-partnered Grok 4.5 aimed at difficult, long-running professional work in July 2026. That Cursor collaboration made coding a central proving ground rather than a secondary use case.
The August Grok 4.6 release used the same $2-per-million input and $6-per-million output token rates. Grok 4.7 keeps that price point while xAI says it improves self-verification and longer-context handling, making the relevant comparison the cost of a completed, checked task rather than raw token price alone.
First-order effects
Developers building coding and knowledge-work workflows can access Grok 4.7 at the prior Grok 4.6 API rates while evaluating xAI’s claims of stronger self-checking and longer-horizon execution.
For SpaceXAI, the release reinforces Grok’s positioning as a lower-cost, faster workhorse for agentic coding, a framing Elon Musk used in public reaction.
Second-order effects
Anthropic and OpenAI face a concrete price-performance reference point for buyers whose workloads require long context and repeated verification, increasing pressure to demonstrate lower cost per completed task rather than benchmark performance alone.
Cursor and other coding-tool customers gain another model option for long-running workflows, making routing decisions more sensitive to reliability across an entire task than to the cost of an individual inference.
Third-order effects
If model providers keep improving verification without raising token rates, agent economics will increasingly be measured by cost per verified task: fewer retries and corrections can matter more than a lower headline input price.
The frontier-model market is moving toward rapid, incremental releases tuned for production workflows, where sustained context, speed, and output quality jointly determine whether an API displaces higher-priced alternatives.
The trend: Frontier AI competition is shifting from standalone model scores toward agentic unit economics: the price of reliably finishing longer, multi-step work.
Grok 4.7 places @SpaceXAI as third, after Anthropic & OpenAI, for agentic coding. When factoring in that Grok is significantly faster & lower cost, it's a great choice for your everyday workhorse.
@spacexai
@spacexai on x
Grok 4.7 is here. It's a notable improvement over Grok 4.6 at the same price and speed.
@elonmusk
Elon Musk on x
Grok 4.7 is a strong combination of intelligence, speed & low cost
@theo
@theo on x
Looking like Grok 4.7 is not quite as cheap as before. Comes out more expensive than Astra on Artificial Analysis 🙃
@haider1
Haider on x
damn... grok 4.7 is a pretty meaningful jump for a “.1” release it uses a new, larger base model than grok 4.6, which means this isn't just more post-training on 4.6
@adamholtererer
Adam Holter on x
Grok is in a perpetual position of “we'll totally beat you next time”
@beffjezos
@beffjezos on x
Better than Fable at Electrical Engineering! Those SpaceX / Tesla datasets giving Grok a huge edge
@grokinsider
@grokinsider on x
Grok 4.7 is out. • Larger base than 4.6, longer RL on hard multi-hour tasks • Better at checking its own work and long context • Same price as 4.6: $2 / $6 per million tokens • CursorBench 4.0: 46.3% (xhigh), up from 4.6's 40.4% • Terminal-Bench 4.0: 38.0%, almost double 4.6 • Ne…
@aksheyd
@aksheyd on x
Excited to bring 4.7 to you all! Numerics aside, it's incredibly capable in Grok Build/Cursor. We spent a lot of hours on the harness, iterating with some of the greatest engineers in the world. Also, the fast mode has INSANE tps. Happy Grok Building :) Lmk what you think!
@teortaxestex
@teortaxestex on x
it has its strengths, I'm sure. Grok 4.6 wasn't a terrible model... But I am happy about the continuing, multi-year vindication of the idea that compute trumps everything. xAI has money, hardware, power, data, and the CEO's legendary Faustian will. And it's not enough. Thanks Elo…
@luminabench
Lumina on x
🚨 Grok 4.7 benchmarks look decent It's a good upgrade over Sol in most areas and has kept the same price: · $2/M input · $6/M output Pretty solid upgrade tbf, just happy to finally have it
@box
@box on x
We put Grok 4.7 from @SpaceXAI to work on a $2 million insurance claim where one deductible error alone changes the calculation by $143,000. In this Box Agent preview, @grok reconciles the claim against the policy and supporting records. It catches an $82,000 duplicate invoice …
@mntruell
Michael Truell on x
Grok 4.7 is out! Smarter and more capable, at the same speed and cost.
@ericzakariasson
Eric Zakariasson on x
grok 4.7 is here, and its our best model so far! try it out in cursor, grok build, api or anywhere you get your tokens! curious to hear what you think here's grok 4.6 vs 4.7 building age of empires ii
@nvidia
@nvidia on x
Grok 4.7 has landed. 🚀 Congrats to @SpaceXAI on its most capable model yet for coding and knowledge work. Proud to support the team with NVIDIA accelerated computing.
@haider1
Haider on x
this is absurd actually CursorBench 4.0: grok 4.7 xhigh: 46.3% — $6.01/task fable 5.1 medium: 46.8% — $7.05/task gpt-5.6 sol max: 41.7% — $8.23/task and compared to grok 4.6 xhigh, it jumped from 41.4% → 46.3% at basically the same cost
@synthwavedd
Leo on x
Well, they launched it, and as expected... Grok 4.7 is pretty terrible. Better luck next time
@louszbd
Lou on x
Really bullish on tasks like EEBench, and Grok 4.7 is looking great here. Amazing work!
@valsai
@valsai on x
We evaluated Grok 4.7 across the Vals benchmark suite. It ranks #24 on the Vals Index at 54.2%, down 5.0 points from Grok 4.6 (#14, 59.2%), but still ahead of Grok 4.5 (#30, 51.5%). Grok 4.7 improves the most on legal and medical work.
@spacexai
@spacexai on x
Grok 4.7 works longer on difficult tasks, checks its work more carefully, and comes with our strongest safeguards to date.
@artificialanlys
@artificialanlys on x
Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index to bring SpaceXAI into the top 4 AI labs. Coding Agent Index performance has also improved, overtaking GPT-5.6 Sol Grok 4.7 scores +2 points over Grok 4.6 on the Intelligence Index, with strong performance on agenti…