Grok✓ ReviewedAnnounced 12 August 2026Reviewed 22 October 2026Grok 4.6 was a step in xAI's model line. Test a migration against today's options.
Reviewed editorial Update
Grok 4.6 was a step in xAI's model line. Test a migration against today's options.
xAI launched Grok 4.6 on 12 August 2026 for long-running agents and interactive or visual work. It remains listed, but Grok 4.7 is the newer flagship. Compare the exact model, pricing tier and reviewed result before changing a working setup.
What changed
xAI launched Grok 4.6 on 12 August 2026, describing a focus on long-running agents and interactive or visual work. Its current global standard API listing has a 500,000-token context window and rates of US$2 input, US$0.50 cached input and US$6 output per million tokens below 200,000 prompt tokens. At 200,000 prompt tokens or more, the listed rates are US$4, US$1 and US$12 respectively for the whole request. Grok 4.5 has the same headline input/output rates and context, but lower cached-input rates. xAI launched Grok 4.7 on 21 September 2026 and now names it its flagship; Grok 4.6 remains listed.
Why it matters
The model number does not settle the migration decision. Grok 4.6 and 4.5 share headline input and output rates, while cached input costs more on 4.6. Long-context and service-route charges also matter. xAI's claims about sustained work are a reason to run a matched test, not proof of a better result for your workflow.
Bit’s takeaway
The human take
A person decides whether the observed result justifies migration and checks the model identity, context tier and bill.
Turn this into a workflow
Bring your task to Bit.
Bit reevaluates the recorded facts when you ask. Historical recommendations do not approve a current action.
Ask Bit about my taskSource-backed facts
- xAI launched Grok 4.6 on 12 August 2026, describing a focus on long-running agents and interactive or visual work. Its current global standard API listing has a 500,000-token context window and rates of US$2 input, US$0.50 cached input and US$6 output per million tokens below 200,000 prompt tokens. At 200,000 prompt tokens or more, the listed rates are US$4, US$1 and US$12 respectively for the whole request. Grok 4.5 has the same headline input/output rates and context, but lower cached-input rates. xAI launched Grok 4.7 on 21 September 2026 and now names it its flagship; Grok 4.6 remains listed.
- xAI news — Grok 4.6 launch announcement ↗
- xAI developer documentation ↗
- xAI: API pricing ↗
- xAI news — Grok 4.7 launch announcement ↗
ReviewedTechnical record
[grok-46-fact; official_fact; product_fact; reviewed 2026-10-22; sources: xAI news — Grok 4.6 launch announcement (https://x.ai/news/grok-4-6); xAI developer documentation (https://docs.x.ai/developers/models); xAI: API pricing (https://docs.x.ai/developers/pricing); xAI news — Grok 4.7 launch announcement (https://x.ai/news/grok-4-7)] xAI launched Grok 4.6 on 12 August 2026, describing a focus on long-running agents and interactive or visual work. Its current global standard API listing has a 500,000-token context window and rates of US$2 input, US$0.50 cached input and US$6 output per million tokens below 200,000 prompt tokens. At 200,000 prompt tokens or more, the listed rates are US$4, US$1 and US$12 respectively for the whole request. Grok 4.5 has the same headline input/output rates and context, but lower cached-input rates. xAI launched Grok 4.7 on 21 September 2026 and now names it its flagship; Grok 4.6 remains listed.
Sources and method