Grok 4.6

Why teams choose it

Tradeoffs to know

• Published benchmark results are vendor-reported and depend on reasoning effort and agent harness.
• Long contexts and xhigh reasoning can increase latency and cost; measure complete task economics rather than token price alone.
• Web, X search, code execution, and function tools require application-level permission and output validation.

When not to use this

• Not ideal for simple tasks where cheaper models in the same lineup are good enough.
• Avoid for latency-sensitive real-time chat when raw response speed outweighs reasoning depth.
• Confirm limits, pricing, and regional availability on the provider side before committing production workloads.

Technical specs

Inputs

text, image

Outputs

text

Capabilities

coding, long-running agents, knowledge work, visual reasoning, function calling, web search, X search, code execution, configurable reasoning

License

Proprietary API

Model string

grok-4-6

Benchmarks

{
  "source": "https://x.ai/news/grok-4-6",
  "deep_swe_v1_1": 65.9,
  "vendor_reported": true,
  "cursorbench_v3_2": 69.9,
  "aa_intelligence_index": 61,
  "frontier_code_v1_1_extended": 61.3
}