Grok 4.6
⧉Free for a limited time: Grok 4.6 is the reasoning model xAI released in August 2026, and NezhaGate gives every account a daily free quota; once it is used up you can switch to the paid grok-4.7. According to xAI it has a 500K-token context window, text and image input, and four reasoning-effort levels (low / medium / high / xhigh); the official list price is $2 / $6 per 1M tokens. It is called like every other model: OpenAI-compatible, with model set to grok-4.6.
Live Test · Playground
Try out Grok 4.6 right here (available after login).
Input
Advanced
Conversation
About Grok 4.6
Grok 4.6 is the reasoning model xAI released in August 2026, offered on NezhaGate through the OpenAI-compatible API free for a limited time: every account gets a daily free quota, and once it is used up you can switch to the paid grok-4.7. According to xAI it has a 500K-token context window, text and image input, and four reasoning-effort levels (low / medium / high / xhigh); the official list price is $2 / $6 per 1M tokens. It is called like every other model: set model to grok-4.6.
Use cases
Free for a limited time, so you can wire up Grok 4.6 and compare results before deciding on a paid model.
xAI's reasoning model, with four official reasoning-effort levels (low / medium / high / xhigh) for problems that need several steps.
xAI quotes a 500K-token context window, which suits summaries and Q&A over long reports and contracts.
xAI lists text and image input, so you can ask about a screenshot or a chart directly.
How to choose
Choose Grok 4.6 to try Grok at no cost; choose grok-4.7 when you need the paid flagship without the daily free quota. Both use the same API, so switching is just the model field.
FAQ
Is Grok 4.6 really free?
How much can I use for free each day?
How is it different from grok-4.7?
How do I call it?
Am I charged for failed calls?
Related Models
Explore other models you can integrate.
Grok 4.7
xAI's flagship reasoning model - long context
Kimi K3
Moonshot's new flagship - long context and agents
DeepSeek V4.1 Flash
DeepSeek's current Flash model - thinking on or off
Qwen3.8 Flash
The light, fast Qwen3.8 - high volume, low cost
Gemini 3.6 Flash
Next-generation fast thinking model · four thinking budgets