Grok AI — All Models Available: catalog, access tiers, pricing, and use cases.

Grok AI has expanded from a single edgy chatbot into a small portfolio of frontier and fast models reachable through the xAI API, the Grok website/app (SuperGrok), and X Premium tiers. The current lineup centers on Grok 4 as the flagship reasoning model, supported by Grok 4 Fast variants for lower-cost throughput, a Grok 4 Heavy configuration available to high-end subscribers, a dedicated Grok Code Fast 1 for agentic coding, and legacy Grok 3 for compatibility. Access depends on your subscription path: X Premium+ unlocks Grok 4 for consumers, SuperGrok subscriptions add higher limits, and SuperGrok Heavy exposes Grok 4 Heavy for the top performance envelope; all are also addressable via xAI’s public API with usage-based pricing.

·····

.....

How the Grok model catalog is organized.

xAI positions the Grok lineup across reasoning, cost-optimized fast inference, and specialized coding. The general rule: Grok 4 is the highest capability model for reasoning and tool use; Grok 4 Fast targets scale and cost; Grok Code Fast 1 is tuned for code generation and agentic coding workflows; Grok 3 remains available for continuity. xAI publicly documents these models and aliases in the Models & Pricing and API docs, together with guidance on context windows and token pricing.

·····

.....

The flagship Grok 4 model and the Heavy configuration.

Grok 4 is xAI’s frontier model advertised for advanced reasoning, native tool use, and real-time search. It is available via X Premium+, SuperGrok on grok.com, and the xAI API. For power users, xAI additionally offers Grok 4 Heavy through the SuperGrok Heavy tier, described as the most powerful configuration of Grok 4, with substantially higher rate limits.

API pricing & capacity: the docs list grok-4 with a large context window and pricing at approximately $3 per million input tokens and $15 per million output tokens, with discounted cached input. Exact metering depends on region and request size; the API page provides the latest details.

·····

.....

What Grok 4 Fast adds for cost-efficient scale.

To support heavier workloads, xAI introduced Grok 4 Fast in two API variants:

• grok-4-fast-reasoning

• grok-4-fast-non-reasoning

Both aim to reduce cost per token and raise throughput versus Grok 4 while preserving useful reasoning capabilities. Pricing starts at roughly $0.20 per million input tokens and $0.50 per million output tokens, with cached input even cheaper. These models are intended for bulk Q&A, large batch summarization, and agent pipelines that need predictable spend and high request volume.

·····

.....

How coding tasks are served by Grok Code Fast 1.

Grok Code Fast 1 is a lightning-fast coding-oriented model designed for agentic coding loops, refactors, and test generation. It exposes a generous context window and budget-friendly pricing brackets in the API, making it a fit for continuous integration bots, LLM code review, and sweeping repository edits. For day-to-day developer chat, Grok 4 still offers the strongest reasoning and tool use, but Code Fast 1 can optimize throughput and latency in automation.

·····

.....

Access paths and which model you actually use.

Grok’s model selection depends on where you log in and what you pay:

• X Premium+: Consumer access to Grok 4, including ad-free X features and higher AI limits compared with lower X tiers.

• SuperGrok (grok.com / app): Direct access to Grok 4 with higher usage limits than baseline consumer tiers.

• SuperGrok Heavy: Access to Grok 4 Heavy and much higher rate limits, positioned for power users and intensive workloads.

• xAI API: Programmatic access to grok-4, grok-4-fast-reasoning, grok-4-fast-non-reasoning, grok-code-fast-1, and grok-3, billed by tokens.

xAI also documents model aliases (e.g., grok-4, grok-4-latest) to help teams pin to a stable release or roll forward automatically—useful for production workloads that balance consistency vs. freshness.

·····

.....

Table — Grok models at a glance.

Model	Category	Context window (headline)	API pricing (indicative)	Primary use cases	Where available
Grok 4	Frontier reasoning	Large (hundreds of thousands)	~$3/M input, $15/M output; cached discounted	Advanced reasoning, tool use, RAG, analysis	X Premium+, SuperGrok, API
Grok 4 Heavy	Frontier (max capacity)	Very large / higher limits	Subscription via SuperGrok Heavy	Highest caps, long sessions, intensive workloads	SuperGrok Heavy
Grok 4 Fast (reasoning)	Cost-efficient reasoning	Up to ~2M tokens (tiered)	from $0.20/M in, $0.50/M out	Batch Q&A, summarization, agent pipelines	API only
Grok 4 Fast (non-reasoning)	High-throughput non-CoT	Up to ~2M tokens (tiered)	similar to reasoning variant	Lightweight tasks, classification, routing	API only
Grok Code Fast 1	Coding / agentic code	Up to ~256k tokens	bracketed per docs	Repo edits, refactors, test generation	API only
Grok 3	Legacy / compatibility	100k+ class	listed alongside Grok 4	Backward-compatible workloads	API, some consumer paths

Notes: Context windows and per-million token rates are summarized from xAI’s docs; always verify the exact figures in the API docs for your region and request size.

·····

.....

Table — Consumer and developer access tiers.

Access path	What you get	Typical model	Who it fits
X Free / lower tiers	Limited Grok access (caps vary by period)	Often constrained	Casual tests on X; not for sustained workloads.
X Premium	Broader features; Grok access varies	Mixed / legacy	Users upgrading from free; check your tier’s AI limits.
X Premium+	Grok 4 + ad-free X + higher limits	Grok 4	Power users who live in X’s ecosystem.
SuperGrok	Website/app login; higher limits	Grok 4	Heavy individual use outside X UI.
SuperGrok Heavy	Highest limits; Grok 4 Heavy	Grok 4 Heavy	Intensive workloads willing to pay premium.
xAI API	Programmatic models, usage billing	All listed API SKUs	Teams, products, and automations at scale.

·····

.....

Where Grok fits in enterprise and public-sector use.

Beyond consumer subscriptions, Grok models are being positioned for enterprise and government adoption with discounted or nominal licensing in specific programs, and availability of Grok 4 and Grok 4 Fast in those channels. If you operate in regulated environments, check whether your deployment requires auditing, data-boundary assurances, and higher quotas—these are increasingly referenced in public agreements.

·····

.....

Practical selection guidance for teams and individuals.

• For highest single-query quality and tool use, choose Grok 4; for the largest personal limits, consider SuperGrok (and SuperGrok Heavy if you truly need the elevated caps).

• For cost-controlled scale—bulk summarization, routing, lightweight agents—use Grok 4 Fast (reasoning or non-reasoning) through the API to leverage the lowest token rates.

• For developer automation and agentic code tasks, wire Grok Code Fast 1 into CI jobs or repo assistants; keep Grok 4 for complex planning/reflection steps.

• If you’re on X, verify your Premium tier—Premium+ grants Grok 4 access; otherwise, consider SuperGrok or the API when your use exceeds consumer limits.

The net effect is a two-lane strategy: a frontier lane for maximum capability (Grok 4 / Heavy) and a fast lane for economical throughput (Grok 4 Fast / Code Fast), with clear subscription and API paths that let you tune for quality, limits, or spend as your workload demands.

.....

FOLLOW US FOR MORE.

DATA STUDIOS

.....[datastudios.org]