xAI's Grok 4.6 Hits 61 on AA Intelligence Index, Lands on Amazon Bedrock

xAI updated its flagship model to Grok 4.6 in just over a month, a rapid cadence that pushed its Artificial Analysis Intelligence Index score to 61 — a mark previously reached only by Anthropic-class models. Grok 4.6 offers a 500,000-token context window, with sibling versions reaching 1M-2M tokens, positioning it as a serious frontier competitor on both intelligence and context length.

The distribution news is equally significant: Grok 4.6 became generally available on Amazon Bedrock, letting AWS enterprise customers use it alongside OpenAI and Anthropic models without switching cloud providers or negotiating separate access. That integration dramatically expands Grok's enterprise reach and reinforces AWS's strategy of being the neutral platform for frontier models — this week's Bedrock additions also included OpenAI Daybreak and GPT-5.6.

Competitively, developers are now weighing Grok 4.6's 61-point intelligence score against cheaper coding workhorses like Gemini 3.7 Flash ($0.75/$3.75 per million tokens), framing a cost-versus-capability tradeoff. Grok's rapid monthly iteration mirrors the model-version-fatigue theme running through the week, as every major lab compresses release cycles.

The skeptical and reputational overhang is severe: xAI faces intense condemnation over a Grok CSAM lawsuit, with attorneys and federal filings arguing the company made deliberate corporate choices to monetize a foreseeable abuse vector that competitors' models refuse to enable. That legal cloud complicates enterprise adoption even as the Bedrock availability lowers the technical barrier — buyers will weigh benchmark strength against brand and liability risk before committing.