Grok 4.6 scores 61 at $2/$6 per million tokens, over 60% cheaper than Opus 5 and GPT-5.6 Sol, and finishes agentic tasks in about half the turns of Opus 5.
Key takeaways · AI-distilled
Artificial Analysis reports a GDPval-AA v2 score of 1753 Elo for Grok 4.6. Its confidence interval overlaps those of Claude Fable 5 and Qwen3.8 Max, limiting claims that the point estimate establishes a clear performance difference.
Grok 4.6's headline input and output prices stayed unchanged from Grok 4.5, but its cache-hit price rose from $0.30 to $0.50 per million tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition →. Cached workloads therefore face a different pricing change.
On AA-Briefcase, Artificial Analysis reports that Grok 4.6 used about half as many turns as Opus 5 max, while scoring below the Opus 5 family. Efficiency gains and task quality remain separate dimensions of that comparison.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
