Grok 4.6
Language modelPrevious generationCard summarizedSpaceXAI's frontier model for coding, agentic tasks, and knowledge work.
Active (superseded by Grok 4.7)● Re-checked 12 Oct 2026 · 18 facts · 0 correctionsSource: Model Card: Grok 4.6 (revision 2026-08-17) (42 pp)
Use it via API
Where to call it
- xAI API
grok-4.6 - Amazon Bedrock
— - Google Vertex AI
— - Microsoft Foundry / Azure
—
Supports
- Image input
- Structured outputs
- Reasoning control
- Tool calling
- Web search
- Prompt caching
Good for
- Long-running agentic coding (CursorBench, DeepSWE, SWE-Marathon, Terminal-Bench)
- Professional knowledge work: documents, spreadsheets, presentations, legal work
- Physical-world engineering: electrical/chip design, procedural 3D and parametric CAD
Watch out for
Overview
Grok 4.6
SpaceXAI's frontier model for coding, agentic tasks, and knowledge work.
- Cost
- $2.00 in
- Context
- 500K
- Input / output
- $2.00 / $6.00
- Max output
- —
- Knows up to
- January 2026
- Released
- 12 Aug 2026
Grok 4.6 extends Grok 4.5 and is developed with Cursor. xAI says it completes longer, harder coding, engineering and office tasks with more autonomy, and reaches results with fewer steps and output tokens than other frontier models. It was xAI's most capable model when released. §1, p.5
grok-4.6- Input / output per 1M
- $2.00 / $6.00
- Context window
- 500K
- Max output
- —
- Knows up to
- January 2026
- Reasoning
- Configurable (low, medium, high, xhigh)
- Weights
- Closed
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses xAI (SpaceXAI)'s Model Card: Grok 4.6 (revision 2026-08-17). Every claim carries its source page.