Grok 4.6 is xAI’s current flagship API model for coding, agentic tool use and knowledge work. It keeps a 500,000-token context window and configurable reasoning, and xAI now points general and code workloads at this SKU rather than Grok 4.5.
It is a hosted commercial model. Headline token prices match the previous flagship, so the upgrade case is capability and platform defaults, not a cheaper rate card.
At a glance
| Fact | Detail |
| --- | --- |
| Provider | xAI |
| Type | Commercial |
| Best for | Coding, agents, research, real-time information workflows |
| Context window | 500,000 tokens |
| API available | Yes |
| Local use | No |
| Last checked | 31 August 2026 |
Our verdict
Use Grok 4.6 when you want xAI’s current general-purpose model for code and agents at $2 / $6 per million tokens. The 500k window is smaller than several million-token rivals, and cached-input rates are not published as a separate public meter on the model page.
| Score | Rating |
| --- | --- |
| Overall | ★★★★☆ |
| Coding | ★★★★☆ |
| Agents | ★★★★☆ |
| Research | ★★★★☆ |
| Budget | ★★★★☆ |
Pricing
Standard API prices per million tokens from xAI’s Grok 4.6 documentation:
| Meter | Price |
| --- | --- |
| Input | $2.00 / million tokens |
| Cached input | Not separately listed on the public model page |
| Output | $6.00 / million tokens |
xAI also lists agent, speech and media meters separately from these text-token rates.
Best for
- Coding and tool-using agents
- Long research packs within a 500,000-token window
- Workflows that use xAI search tools for current information
Avoid when
- You need local inference or open weights
- You need a million-token context window
- Your workload is simple bulk classification
What changed
- 31 August 2026: Profile created for Grok 4.6. Context window, $2 / $6 token rates and flagship status checked against xAI documentation.
Related news
- Grok 4.5: pricing, capabilities and latest changes
- SpaceXAI ships Grok 4.5, Musk calls it an Opus-Class model at half the price
- OpenAI’s GPT-5.6 push is really a pricing war for enterprise AI
