Cafiyn Pulse
xAI · Verified 2026-09-05

Grok 4.6 pricing, and the long-context penalty.

$2 and $6 per million tokens, which undercuts most of the mid tier. Both rates double above 200k tokens.

Grok 4.6
xAI · premium tier
Verify against xAI's pricing page ↗
$2.00
per 1M input tokens
$6.00
per 1M output tokens

Worth knowingRate doubles on prompts of 200k tokens or more

A typical 1,200-in / 400-out token request costs about $0.0048, or roughly $4.80 per 1,000 requests.

Grok 4.6 runs $2 per million input tokens and $6 per million output, which xAI describes as its most capable model. That output price is notably low for the tier: Claude Sonnet 5 is $10 and GPT-5 is $10 on output, so Grok is roughly 40 percent cheaper on generation-heavy work while matching Sonnet 5 on input.

The catch is context length. xAI doubles both input and output rates on prompts of 200k tokens or more, taking Grok 4.6 to $4 and $12. That is a cliff rather than a gradient, so a workload that occasionally crosses 200k pays double on those calls with no warning in the headline price. If you run long-context work, budget at the higher tier or keep prompts deliberately under the threshold.

Best for
  • Generation-heavy workloads where the low output price compounds
  • Teams already using X or xAI tooling, where ecosystem fit outweighs marginal quality
  • Short and medium context work that stays comfortably under 200k tokens
Head to head

Grok 4.6 vs the alternatives.

ModelInput / 1MOutput / 1MSample request
Grok 4.6$2.00$6.00$0.0048
Claude Sonnet 5$2.00$10.00$0.0064
GPT-5$1.25$10.00$0.0055
Sample request assumes 1,200 input and 400 output tokens, a typical chat exchange. Your actual ratio will differ, run your own numbers in the calculator below.
FAQ

Common questions.

How much does Grok 4.6 cost?

$2 per million input tokens and $6 per million output tokens for prompts under 200k tokens. At 200k tokens or more both rates double, to $4 and $12.

Is Grok 4.6 cheaper than Claude Sonnet 5?

On output yes, substantially: $6 against $10 per million tokens. Input is identical at $2. On generation-heavy workloads Grok comes out meaningfully cheaper, provided your prompts stay under the 200k-token threshold where the rate doubles.

What happened to Grok 4?

Grok 4 is no longer listed on xAI's models page. It was $3 input and $15 output when last verified in July 2026, so the current Grok 4.6 at $2 and $6 is a large reduction on both legs.

More models

Browse other pricing pages.

Open the tool.

Live math against xAI's current pricing plus every other provider in the roster.

Compare Grok 4.6 on my usage