SpaceXAI has released Grok 4.6, scoring 61 on the Artificial Analysis Intelligence Index. That puts it above Moonshot's Kimi K3, tied with OpenAI's GPT-5.6 Sol Max, and five points clear of its predecessor Grok 4.5 High. Anthropic's Claude Opus 5 and Fable 5 still hold the top two spots.
The pricing is the real story for enterprise buyers. Grok 4.6 costs $2 per million input tokens and $6 per million output tokens for prompts under 200K tokens, totaling $8 per million. That is less than half the $35 standard-mode cost of GPT-5.6 Sol and a fraction of Claude Opus 5's $30 per million output price. The model targets long-running agents, coding, and knowledge work, and SpaceXAI is doubling included usage in both Grok Build and Cursor for the first week of launch.
Read the full piece for the complete pricing comparison table across 25 frontier models, from Meta's Muse Spark at $0.30 per million total to GPT-5.6 Sol Fast at $70 per million. The table makes the value positioning of Grok 4.6 concrete and puts the benchmark gains in cost-per-task context that benchmark scores alone do not capture.
[READ ORIGINAL →]