Elon Musk Said Grok 4.7 Would Exceed Every Model. Did It Deliver?
SpaceXAI shipped Grok 4.7, and impartial scores place it fourth amongst frontier AI labs.
Artificial Analysis, an impartial benchmarking agency, scored the mannequin at 46 on its Intelligence Index, 2 factors above Grok 4.6. Claude Fable 5.1, GPT-6 Astra, and Claude Opus 5 all rank increased.
Grok 4.7 Pushes SpaceXAI Into the Top Four AI Labs
Elon Musk said before launch that the mannequin would run on a 2.1-trillion-parameter base and exceed all present fashions. He added that the mannequin coaching included SpaceX firm information.
Follow us on X to get the newest information because it occurs
But how does Grok 4.7 really carry out? On AA-Briefcase, an Artificial Analysis benchmark for life like skilled work duties, Grok 4.7 scored 1657 Elo. That marks an 111-point soar over Grok 4.6 and brings the mannequin’s stage near Claude Opus 5.
On GDPval-AA, the mannequin reached 1695 Elo, which places it 90 factors forward of its predecessor. Coding outcomes moved in the identical course.
Paired with Grok Build, its personal coding agent, Grok 4.7 scored 56 on the Coding Agent Index. That is 9 factors above Grok 4.6.
The mannequin handed GPT-5.6 Sol and now trails solely Anthropic’s Claude Fable 5.1, GPT-6 Astra, and Opus 5.
“Grok 4.7 scores 46 on the Artificial Analysis Intelligence Index to deliver SpaceXAI into the highest 4 AI labs. Coding Agent Index efficiency has additionally improved, overtaking GPT-5.6 Sol,” the put up learn.
Elsewhere, the mannequin barely moved. It gained 4.5 share factors on Terminal-Bench 4.0 and three factors on GDP.pdf. Scores slipped on the AA-LCR and AutomationBench-AA checks. Grok 4.5 topped AutomationBench in July, which backed Musk’s declare that it matched Claude Opus.
The Gains Come With a Token Bill
The charge card didn’t change. Grok 4.7 nonetheless prices $2 per million enter tokens and $6 per million output tokens. Cache hits keep discounted to $0.50, and the five hundred,000-token context window carries over from Grok 4.6.
The mannequin does significantly extra work for every reply, nevertheless. Artificial Analysis measured roughly 81,000 output tokens per Index activity, towards 36,000 for Grok 4.6 and 27,000 for GPT-6 Astra. Therefore, regular per-token pricing nonetheless produces a better invoice per activity.
That urge for food for tokens lands on a division that has already reported a $1.26 billion loss. Meanwhile, Musk mentioned on September 14 that Grok 4.8 would end coaching inside every week.
He additionally expects Grok 5 to be the model that reaches synthetic common intelligence. The subsequent launch will check whether or not SpaceXAI can climb the rankings with out burning further compute to take action.
Subscribe to our YouTube channel to observe leaders and journalists present skilled insights
The put up Elon Musk Said Grok 4.7 Would Exceed Every Model. Did It Deliver? appeared first on BeInCrypto.
