Grok 4.6 is xAI's flagship large language model, built for developers and teams running long-horizon agents, coding workflows, and knowledge work. Released on August 12, 2026, it is the default model inside Grok Build and the coding agent, and it is available through the xAI API, Cursor, OpenRouter, Vercel, and Cloudflare. The model carries roughly 1.5 trillion parameters and runs at about 80 tokens per second.
Core Features
- 500,000-token context window with configurable reasoning effort (low, medium, high, xhigh).
- Multimodal input: accepts text and images, returns text, with no output token cap.
- Native tool use: function calling, web search, X search, and server-side code execution.
- Long agent-loop support via prompt cache keys and context compaction for multi-turn tasks.
- Scored 1753 Elo on GDPVal-AA v2 and 65.9% on DeepSWE v1.1 for coding agents.
Use Cases
- Autonomous coding agents that plan, edit, and test across large codebases.
- Long-running research and document-analysis assistants that keep full context.
- Multi-step product prototyping where visual and interactive first passes matter.
Pricing
API pricing starts at $2.00 per million input tokens and $6.00 per million output tokens. Cached input is $0.50 per million. Prompts of 200k tokens or more bill at $4 / $1 / $12 per million. A 2x Priority Processing lane is available. Try it free inside Grok Build.
Our Take
Best for engineering teams that need a long-context, agent-ready model without leaving the xAI or Cursor stack. The trade-off is that once a prompt crosses 200k tokens, the entire request bills at the higher rate, so cache keys are essential to control cost.
Explore more AI models and engines on aifreetool.site.
