Grok 4.6
Grok 4.6 is SpaceXAI's August 12, 2026 flagship model with the API ID grok-4.6, a 500,000-token context window, text and image input, text output, no documented text-output limit, and base pricing of $2 input, $0.50 cached input, and $6 output per million tokens below 200,000 prompt tokens.
Grok 4.6 is a current frontier-value candidate rather than a minor Grok 4.5 refresh. It improves long-running agents and visual or interactive work, is available across the SpaceXAI API, Grok Build, Cursor, GitHub Copilot, and model gateways, and reaches a strong independent intelligence score at substantially lower token prices than Fable 5.1 or Astra.
SpaceXAI's announcement, model page, and release notes control the identity, limits, pricing, reasoning, tools, and availability. Launch benchmark tables remain vendor-reported even when they cite public results. Artificial Analysis provides the independent comparison configuration used on the dated frontier-model page.
- Run long coding and repository agents across SpaceXAI, Cursor, or GitHub Copilot surfaces.
- Build interactive and visual applications through repeated tool-using loops.
- Route knowledge-work tasks that need frontier-class capability at moderate token prices.
- Use web, X search, and code execution where the SpaceXAI tool surface fits the workload.
The API model ID is grok-4.6. SpaceXAI documents text and image input, text output, 500,000 context tokens, a January 2026 knowledge cutoff, no text-output limit, and low, medium, high, or xhigh reasoning effort. Responses and Chat Completions APIs support function calling, web search, X search, and code execution.
Below 200,000 prompt tokens, the model costs $2 input, $0.50 cached input, and $6 output per million tokens. Above that threshold the rates double to $4, $1, and $12. A fast variant costs twice the applicable rate, so long prompts and latency routing need explicit cost controls.
Artificial Analysis measured Grok 4.6 high at 61 on its current Intelligence Index, tying several frontier configurations while keeping token prices well below Fable 5.1 and Astra. That does not make it universally best: Fable scores higher at max effort, Gemini is cheaper and faster, and provider-specific tools or governance can dominate the decision.
Grok 4.6 FAQ
Common questions about Grok 4.6.
What is the Grok 4.6 API model ID?+
Use grok-4.6 in the SpaceXAI API. The model supports the Responses and Chat Completions APIs, with low, medium, high, and xhigh reasoning effort.
How much does Grok 4.6 cost?+
Below 200,000 prompt tokens, pricing is $2 input, $0.50 cached input, and $6 output per million tokens. Above 200,000 prompt tokens the three rates double. The fast variant is twice the applicable price.
Is Grok 4.6 better than GPT-6 Astra or Claude Fable 5.1?+
It can be the better value for long-running coding or knowledge work, but it is not the universal capability leader. Fable 5.1 currently leads the selected independent composite at max effort, Astra is strong in OpenAI's coding-agent harness, and Gemini 3.8 Flash costs less. Compare validated task success and total cost.