Grok✓ ReviewedTest carefullyChecked 5 September 2026Grok 4.6 is xAI's new flagship. The upgrade case is your task, not the benchmark.
xAI released Grok 4.6 as the model it now recommends for coding and chat, built on Grok 4.5 with more focus on long-running agents and interactive or visual work. Access, headline pricing and the context window are unchanged, so the newer number is not a reason on its own to migrate a working setup.
Bit’s takeaway
What changed
xAI released Grok 4.6 on 12 August 2026 through the xAI API, Grok Build and Cursor, and its model catalogue now recommends 4.6 for coding and chat, calling it the most intelligent and fastest model it has built. It builds on Grok 4.5 with a particular focus on long-running agents and more ambitious interactive and visual work, and shows more self-testing and verification on extended tasks. Access, the headline input and output prices and the 500,000-token context window match Grok 4.5; cached-input pricing is higher. Grok 4.5 remains listed and available with its original pricing and limits.
Why it matters
A newer model number invites an automatic upgrade, but the parts that actually cost you time are unchanged: the price per token and the tools it reaches are the same as 4.5. The case for moving is narrower and more testable than the launch suggests, it rests on whether the long-horizon and visual-work gains show up on your own task, and the one thing that did move, cached-input pricing, works against a naive switch.
Who should care
- Teams already running a Grok 4.5 coding or agentic workflow
- Developers comparing flagship models on long-running or visual work
What to do
Keep your Grok 4.5 workflow running and put 4.6 beside it on the same long-running or visual task with a fixed source pack and acceptance checklist. Score the reviewed result and the actual bill, including cached-input cost, before you switch anything in production.
The human take
Tools change fast. Your judgment matters more.
A person decides whether the long-horizon and visual-work gains actually change the accepted result, and reads the real bill before migrating off a working 4.5 setup.
Affected guidance
Put this to work
Verified facts
- xAI released Grok 4.6 on 12 August 2026 as its current flagship, recommended for coding and chat and described as its most intelligent and fastest model, building on Grok 4.5 with a focus on long-running agents and interactive or visual work; access, the headline input and output pricing and the 500,000-token context window are unchanged from Grok 4.5, while cached-input pricing is higher.Checked
Sources and method