August 12, 2026
Cheap, smart, and deeply suspicious
SpaceXAI's Grok 4.6 Scores 61 on the Artificial Analysis Intelligence Index
Grok climbs back into the AI big leagues, but the comments are split between bargain-hunters and side-eye
TLDR: Grok 4.6 jumped to a top-tier AI score while keeping its main price lower than several rivals, making it look like a real threat in the chatbot race. But the comments stole the show: some cheered the value, while others questioned who uses it and how it got so good.
SpaceXAI’s Grok 4.6 just posted a 61 on the Artificial Analysis Intelligence Index, putting it basically neck-and-neck with some of the biggest names in AI and only a hair behind Anthropic’s top models. In plain English: Grok has gone from “interesting underdog” to serious contender, and it did it without raising its main price. That’s the part that made the crowd sit up. One commenter practically turned into Grok’s sales team, raving that services like Cursor now get you a frontier-level bargain and saying your subscription stretches much farther than with OpenAI or Anthropic.
But this was not a polite golf clap kind of thread. Oh no. The hottest reaction was pure conspiracy-mode side eye: one user openly wondered if Grok got this good by “stealing” or distilling Anthropic’s models. That spicy accusation instantly gave the whole discussion a reality-show vibe. On the other side were the practical skeptics, including one brutally funny commenter who said they’ve “never met a single human being” using Grok for coding — a line with strong meme energy.
Then came the wallet-watchers. While Grok’s headline price stayed the same, users noticed cache pricing jumped, and one heavy coder complained that this is where 80% of the bill can come from. So the verdict from the internet? Grok 4.6 looks fast, cheap, and suddenly elite on paper — but the comments are torn between deal excitement, trust issues, and “cool score, who actually uses this?” energy.
Key Points
- •Artificial Analysis reports that Grok 4.6 scored 61 on its Intelligence Index, up 5 points from Grok 4.5 and 23 points from Grok 4.3.
- •The article says Grok 4.6 matches GPT-5.6 Sol on the index, trails Claude Opus 5 and Claude Fable 5, and sits just ahead of Kimi K3.
- •On agentic benchmarks, Grok 4.6 records 1753 Elo on GDPval-AA v2, 50.7% on τ³-Banking, and 88.4% on Terminal-Bench v2.1.
- •Grok 4.6 keeps Grok 4.5’s headline pricing at $2 per million input tokens and $6 per million output tokens, while cache-hit pricing rises to $0.50 per million tokens.
- •On AA-Briefcase, Grok 4.6 scores 1577 Elo and is described as more turn- and token-efficient than Claude Opus 5 (max) in long-horizon tasks.