Grok 4.6 scores 61 on Artificial Analysis index at $2/$6 per 1M tokens

Grok 4.6, developed by xAI, scores 61 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Sol and trailing Claude Opus 5 by two points, while priced at $2 per 1M input tokens and $6 per 1M output tokens—over 60% cheaper than Opus 5's $5/$25. The model averages about 53 turns and 0.5B input tokens on long-horizon agentic tasks, versus Opus 5's ~103 turns and ~2.0B, making it a cost-efficient choice for reasoning-heavy and multi-turn tool-use pipelines, though cache-hit pricing rose to $0.5 per 1M tokens from $0.3.

Hacker News https://artificialanalysis.ai/articles/grok-4-6-benchmarks-and-analysis Grok 4.6 scores 61 on Artificial Analysis index at $2/$6 per 1M tokens Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated. Grok 4.6 hits frontier-tier intelligence 61, matching GPT-5.6 Sol, two points behind Claude Opus 5 at $2/$6 per 1M tokens — 60%+ cheaper than Opus 5's $5/$25 — and is dramatically more turn-efficient on long-horizon agentic work, averaging ~53 turns and ~0.5B input tokens versus Opus 5's ~103 turns and ~2.0B. For reasoning-heavy and multi-turn tool-use pipelines, it's now the clear Pareto pick on cost-per-task, though note cache-hit pricing rose to $0.5/1M up from $0.3 , so recompute your prompt-caching economics before switching. Grok 4.6 achieves a score of 61 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Sol's score while maintaining the same $2/$6 per 1M input/output tokens pricing, 60% lower than comparable models like Claude Opus 5; this enables cost-effective deployment of high-intelligence models for agentic knowledge work tasks.