Grok API pricing (2026): the per-token cost table

Short answer: Grok's flagship 4.6 costs $2 input / $6 output per million tokens (500K context). The budget pick is Grok 4.1 Fast at $0.20 / $0.50 with a 2M context window, and coding-specific work runs on Grok Code Fast 1 at $0.20 / $1.50. xAI applies roughly 3x output-to-input pricing on flagship tiers, and doubles rates above 200K tokens of context on the 4.6/4.5/4.3 models.

Grok API model pricing (per 1M tokens)

ModelInputOutputCached inputContext
Grok 4.6 (flagship)$2$6$0.50500K
Grok 4.5$2$6$0.30500K
Grok 4.20$2$62M
Grok 4.3$1.25$2.50$0.201M
Grok 4.1 Fast (cheapest general)$0.20$0.502M
Grok Code Fast 1 (coding)$0.20$1.50$0.02256K
Grok 3 Mini$0.30$0.50128K

Prices are USD per million tokens (MTok). xAI applies higher-context pricing above 200K tokens on the flagship 4.6/4.5/4.3 tiers, effectively doubling the listed rate for requests that push into the top of those context windows — check the current xAI docs before budgeting a workload that regularly runs long-context requests.

Worked example

A request that sends 20,000 input tokens and generates 2,000 output tokens on Grok 4.1 Fast:

The same request on flagship Grok 4.6 costs (20,000 × $2 + 2,000 × $6) / 1,000,000 = $0.052 — roughly 10x Grok 4.1 Fast, for meaningfully stronger reasoning and real-time X data grounding.

How Grok compares to the majors

Grok's positioning is mixed: 4.1 Fast at $0.20 input is competitive with the cheapest tiers from OpenAI and Google, but flagship Grok 4.6 at $2/$6 sits closer to mid-tier pricing than to a true budget option, and DeepSeek undercuts Grok at every comparable tier. Where Grok earns its price is real-time X (Twitter) data access and SuperGrok's consumer app, not raw per-token cost. Full context in our Grok review and ChatGPT vs Grok comparison.

Full cross-provider breakdown: cheapest LLM API, OpenAI pricing, Anthropic pricing, DeepSeek pricing.