Last verified
FRONTIER LABFOUNDED 2023SPACEX SUBSIDIARYMEMPHIS COLOSSUSX / GROK NATIVE

Grok API Pricing

Elon Musk's AI lab, founded March 2023 as X.AI Corp. Ships the Grok family — frontier reasoning and coding models trained on the Memphis Colossus supercluster (150 MW peak, plans for 1M+ GPUs). Acquired by SpaceX on Feb 2, 2026 at a $250B xAI valuation ($1.25T combined entity). Brand still operates as xAI; API still at docs.x.ai; Grok still integrated natively into X (Twitter).

Production models
6
Grok 4.5 + Build + 4.20
Founded
2023
Mar 9 · 3 years
Latest model
Grok 4.5
released Jul 8, 2026
Latest input
$2/M
Grok 4.5 input
Max context
2M tokens
Grok 4.20 Multi-Agent
Cached input
$0.50/M
Grok 4.5 cache
§ 01 / LINEUP

The full roster.

Side-by-side →
FRONTIER · REASONING

Grok 4.20 (0309) Reasoning

Dated 03-09 snapshot of Grok 4.20 in reasoning mode.

Input
$1/M
Output
$3/M
1M ctx · text-only
FRONTIER · REASONING

Grok 4.20 (0309) Non-Reasoning

Dated 03-09 snapshot of Grok 4.20 in non-reasoning (latency-optimised) mode.

Input
$1/M
Output
$3/M
1M ctx · text-only
FRONTIER · REASONING

Grok 4.20 Multi-Agent (0309)

Multi-agent orchestration variant of Grok 4.20 — 2M context window (only Grok variant with 2M).

Input
$1/M
Output
$3/M
2M ctx · text-only
FRONTIER · REASONING

Grok 4.5 LAUNCHED JUL 2026

xAI pricing docs list grok-4.5 at $2.00/M input, $0.50/M cached input, and $6.00/M output.

Input
$2/M
Output
$6/M
500K ctx · vision
FRONTIER · REASONING

Grok 4.3 LAUNCHED MAY 2026

Still listed after Grok 4.5 launch.

Input
$1/M
Output
$3/M
1M ctx · text-only
CODING

Grok Build 0.1 LAUNCHED MAY 2026

xAI's fast agentic-coding model, now listed on the standard Code API pricing table at $1/M input and $2/M output (256K context).

Input
$1/M
Output
$2/M
256K ctx · vision
RETIRED · MAY 2026

Grok 4.1 Fast LAUNCHED NOV 2025

Archive price retained from the verified snapshot: $0.20/M input, $0.05/M cached input, $0.50/M output.

Input
$0.20/M
Output
$0.50/M
256K ctx · text-only
RETIRED · MAY 2026

Grok 4 Fast LAUNCHED SEP 2025

Archive price retained from the verified snapshot: $0.20/M input, $0.05/M cached input, $0.50/M output.

Input
$0.20/M
Output
$0.50/M
2M ctx · text-only
RETIRED · MAY 2026

Grok 4 LAUNCHED JUL 2025

Archive price retained from the verified snapshot: $3/M input, $0.75/M cached input, $15/M output.

Input
$3/M
Output
$15/M
256K ctx · text-only
RETIRED · MAY 2026

Grok 3 LAUNCHED FEB 2025

Archive price retained from the verified snapshot: $3/M input, $0.75/M cached input, $15/M output.

Input
$3/M
Output
$15/M
131K ctx · text-only
RETIRED · SEP 2025

Grok 3 Mini LAUNCHED FEB 2025

Archive price retained from the verified snapshot: $0.30/M input and $0.50/M output.

Input
$0.30/M
Output
$0.50/M
131K ctx · text-only
RETIRED · AUG 2025

Grok 2 (1212) LAUNCHED DEC 2024

Official xAI 1212 launch post announced grok-2-1212 and grok-2-vision-1212 at $2/M input and $10/M output after efficiency improvements.

Input
$2/M
Output
$10/M
131K ctx · text-only
RETIRED · AUG 2025

Grok 2 Vision (1212) LAUNCHED DEC 2024

API ID grok-2-vision-1212.

Input
$2/M
Output
$10/M
32K ctx · vision
RETIRED · MAY 2026

Grok Code Fast 1 LAUNCHED AUG 2025

Archive price retained from the verified snapshot: $0.20/M input, $0.02/M cached input, $1.50/M output.

Input
$0.20/M
Output
$2/M
256K ctx · text-only
RETIRED · FEB 2025

Grok Beta LAUNCHED NOV 2024

API ID grok-beta.

Input
$5/M
Output
$15/M
131K ctx · text-only
RETIRED · FEB 2025

Grok Vision Beta LAUNCHED NOV 2024

API ID grok-vision-beta.

Input
$5/M
Output
$15/M
32K ctx · vision
§ 02 / SHELF

All side-by-side.

Methodology →
Model Input /M Output /M Cached Context Max output Vision Tools Tier
Grok 4.20 (0309) Reasoning $1.25 $2.50 $0.2−84% 1M Frontier
Grok 4.20 (0309) Non-Reasoning $1.25 $2.50 $0.2−84% 1M Frontier
Grok 4.20 Multi-Agent (0309) $1.25 $2.50 $0.2−84% 2M Frontier
Grok 4.5 FLAGSHIP $2.00 $6.00 $0.5−75% 500K Frontier
Grok 4.3 $1.25 $2.50 $0.2−84% 1M Frontier
Grok Build 0.1 $1.00 $2.00 $0.2−80% 256K Active
Grok 4.1 Fast RETIRED Retired May 15, 2026 → grok-4-3 Retired
Grok 4 Fast RETIRED Retired May 15, 2026 → grok-4-3 Retired
Grok 4 RETIRED Retired May 15, 2026 → grok-4-3 Retired
Grok 3 RETIRED Retired May 15, 2026 → grok-4-3 Retired
Grok 3 Mini RETIRED Retired Sep 30, 2025 → grok-4-fast Retired
Grok 2 (1212) RETIRED Retired Aug 14, 2025 → grok-3 Retired
Grok 2 Vision (1212) RETIRED Retired Aug 14, 2025 → grok-3 Retired
Grok Code Fast 1 RETIRED Retired May 15, 2026 → grok-4-3 Retired
Grok Beta RETIRED Retired Feb 28, 2025 → grok-2-1212 Retired
Grok Vision Beta RETIRED Retired Feb 28, 2025 → grok-2-vision-1212 Retired
§ 03 / PRICE CURVE

Pricing across the lineup.

How xAI priced 13 models · NOV 24 → JUL 26.

oldest → newest →
$20$15$10$5 Grok Beta NOV 24 Grok 2 (1212) DEC 24 Grok 3 Mini FEB 25 Grok 4 JUL 25 Grok 4 Fast SEP 25 Grok 4.3 MAY 26 Grok 4.5 JUL 26
Input · newest $2/M
Output · newest $6/M
Each point is a model at its listed $/M price.

The last 12 months at xAI.

Major releases, retirements, and corporate moves. Sourced from xAI press releases, docs.x.ai migration guides, and recent merger filings.

JUL 08 · 2026
Grok 4.5 released on the xAI API — latest coding, agentic, and knowledge-work model at $2/$6, 500K context, configurable reasoning
RELEASE
MAY 19 · 2026
Grok Build 0.1 released — early-access fast coding model for agentic coding at $1/$2
RELEASE
MAY 15 · 2026
Retirement wave — Grok 4, Grok 4 Fast, Grok 4.1 Fast, Grok Code Fast 1, Grok 3 all retired; traffic redirects to grok-4.3
PRICING
MAY 12 · 2026
Grok 4.3 shipped to API — 40% price cut vs Grok 4 ($1.25/$2.50 input/output), 1M context, agentic tool-calling focus
RELEASE
APR · 2026
Michael Nicolls joins as President of xAI
CORPORATE
FEB 02 · 2026
SpaceX acquires xAI at $250B xAI valuation ($1.25T combined entity) — funds orbital data centre build-out
CORPORATE
OCT · 2025
Grokipedia launched — AI-generated Wikipedia alternative
RELEASE
SEP · 2025
500 data annotators laid off; xAI shifts to model-led RLHF instead of manual labelling
CORPORATE
AUG 14 · 2025
Grok 2 family retired — grok-2-1212 + grok-2-vision-1212 shut down, redirect to Grok 3 then Grok 4
PRICING
JUL 09 · 2025
Grok 4 released — first frontier-tier model at $3/$15 per M; Grok Heavy consumer subscription launched at $300/mo
RELEASE
MAR · 2025
Valuation hits $80B after X Corp acquisition; X / Twitter folded into xAI's stack
CORPORATE
FEB 17 · 2025
Grok 3 released — adds reflection / reasoning mode; Grok 3 Mini at $0.30/$0.50
RELEASE
DEC · 2024
Memphis Colossus operational — 122-day build, 150 MW peak draw, plans for 1M+ GPUs
CORPORATE
§ 04 / ACCESS

Where to get it.

Methodology →
§ 04 / BEST FOR

Which xAI for what.

More scenarios →
If you need the latest Grok for coding, agents, and knowledge work
Grok 4.5
Profile →
If you need a fast coding-specialist model
Grok Build 0.1
Profile →
If you need cheaper Grok with 1M context
Grok 4.3
Profile →
If you need the longest context window in the Grok line…
Grok 4.20 Multi-Agent (2M)
Profile →
If you need a pinned weights snapshot for reproducibility…
Grok 4.20 (0309)
Profile →
If you want low-latency Grok without chain-of-thought…
Grok 4.20 (0309) Non-Reasoning
Profile →
If your stack lives on X / Twitter and you want native AI…
Grok via X integration
Open X →
If you need vision / image / audio on a frontier model…
Use Gemini or GPT instead
Compare →
§ 06 / BACKGROUND

The company behind it.

x.ai →

xAI was founded by Elon Musk on March 9, 2023 as X.AI Corp, with 11 founding researchers (including ex-DeepMind's Igor Babuschkin). The pitch: build a maximally truth-seeking AI inside Musk's ecosystem — X (Twitter) for data and distribution, Tesla / SpaceX for compute and capital. CEO: Elon Musk.

The lab moved fast. Grok 1 shipped November 2023 (weights later open-sourced). Grok 2 in August 2024 added image generation. Grok 3 in February 2025 introduced the reflection / reasoning mode that anchors the current line. Grok 4 in July 2025 was the first frontier-tier Grok ($3/$15) and shipped alongside SuperGrok Heavy, a $300/mo consumer tier that exposed early access to the largest variants. Grok 4.3 on May 12, 2026 cut input prices ~40% to $1.25/$2.50 and the next-generation Grok 4.20 family arrived in 03-09 snapshot form.

The defining infrastructure asset is the Memphis Colossus supercluster, brought online in December 2024 after a remarkable 122-day build. Peak draw 150 MW; expansion plans target 1M+ GPUs. This is the reason xAI can offer flagship pricing at $1.25/M input — the per-token economics are subsidised by owning compute rather than renting it.

The corporate structure changed dramatically in 2026. On Feb 2, 2026, SpaceX acquired xAI at a $250B xAI valuation and $1.25T combined-entity valuation — the largest tech merger by valuation on record. Stated rationale: orbital data centres + more capital runway for AI buildout against OpenAI and Anthropic. xAI continues to operate under its own brand and API surface; only the parent changed.

Compared to Anthropic and OpenAI: cheaper per-token at frontier ($1.25/$2.50 vs $5/$30 GPT-5.5), smaller surface (mostly text-focused: Grok Build 0.1 accepts image input for coding workflows, while image generation, video, and voice live in separate Imagine / Voice APIs), tied to X / Twitter brand for better and worse. SWE-bench Verified for Grok 4.3 trails Claude Opus 4.7 by roughly 14pp (~74% vs 87.6%).

§ 07 / COMPETITORS

Other frontier labs.

All providers →
PRIMARY RIVAL

OpenAI

Bigger surface area (image / audio / video / embeddings under one roof), bigger consumer brand. 4× more expensive at flagship input ($5 vs $1.25). xAI wins on per-token cost; OpenAI wins on multimodality and ecosystem.

GPT-5 family
CODING LEADER

Anthropic

Leads SWE-bench Verified at 87.6% vs Grok 4.3's ~74%. Zero-retention defaults and Responsible Scaling Policy. 4× more expensive at flagship ($5 vs $1.25) but stronger on coding agents.

Claude family
CLOUD GIANT

Google

2M context on Gemini 2.5 Pro matches xAI's Grok 4.20 Multi-Agent. Native multimodal (text + image + audio + video). Pricier per token but vastly broader ecosystem via Workspace + Vertex.

Gemini family
OPEN-WEIGHT GIANT

Alibaba (Qwen)

The Qwen (Tongyi Qianwen) model family from Alibaba Cloud's Tongyi Lab. First released in 2023, Qwen is the most-downloaded open-weight model family in the world — most tiers ship under Apache 2.0 on Hugging Face and ModelScope, while the proprietary Max tier is served through Alibaba Cloud's Model Studio.

28 models · from $0.05/M
CHINA CONSUMER LEADER

ByteDance (Doubao)

The Doubao model family from ByteDance — parent of TikTok and Douyin. Served through the Volcano Ark (火山方舟) platform on ByteDance's Volcano Engine cloud, Doubao powers China's most-used consumer AI app and undercuts most frontier labs on price.

14 models · from $0.02/M

Frequently asked.

Practical questions about xAI, the Grok lineup, the SpaceX merger, and how text-only frontier pricing actually shakes out.

Q · 01 Which Grok model should I start with? +
Grok 4.5 for the latest frontier model: coding, agentic tasks, knowledge work, configurable reasoning, and text + image input at $2/$6. Pick Grok 4.3 when you want a cheaper 1M-context Grok option at $1.25/$2.50. Pick Grok 4.20 Multi-Agent if you need the 2M-context window.
Q · 02 Does Grok support vision, audio, or image generation? +
The main Grok chat/reasoning line is still mostly text-focused, but Grok Build 0.1 accepts image input for coding workflows. Image generation and video are handled by the separate Imagine API; voice by the Voice API. For a single frontier text+vision+audio model, compare Gemini, GPT, or Claude.
Q · 03 Did SpaceX really acquire xAI? +
Yes — closed Feb 2, 2026 at a $250B xAI valuation ($1.25T combined entity). The deal was framed as funding for orbital data centres and giving xAI more capital runway against OpenAI / Anthropic infrastructure spend. The xAI brand, console at console.x.ai, and Grok product line all continue to operate as before; only the parent ownership changed.
Q · 04 Does xAI offer EU data residency? +
No — xAI does not publish EU-specific regions or data-residency commitments. All Grok API traffic terminates at xAI's US infrastructure (Memphis + Bay Area). No AWS Bedrock / Azure / Vertex resale of Grok models means there's no cloud-provider workaround. If EU data residency is a hard requirement, pick Claude on Bedrock EU, Gemini on Vertex EU, or Mistral.
Q · 05 What happened in the May 15, 2026 retirement wave? +
xAI retired five models on the same day: Grok 4, Grok 4 Fast, Grok 4.1 Fast, Grok Code Fast 1, and Grok 3. All requests to those slugs now redirect to grok-4.3 and bill at Grok 4.3 pricing ($1.25/$2.50). This consolidated the lineup down to the current 4 active models. Migration guide: docs.x.ai/developers/migration/may-15-retirement.
Q · 06 Can I get a discount? +
The biggest realistic discount on Grok is the cached input rate of $0.20/M (vs $1.25/M base — 84% off). Batch and Flex tiers are not published for Grok the way they are for OpenAI; xAI's pricing page lists only standard + cache. SuperGrok Heavy ($300/mo) is consumer-side, not API. No committed-spend tier is publicly advertised.
Q · 07 How does Grok compare to Claude and GPT? +
Cheaper per token (4× less than GPT-5.5 input, 12× less than Opus 4.5 output). Smaller surface area — text only, no native vision / image gen / audio. Trails on coding-agent benchmarks (~74% SWE-bench Verified vs Opus 4.7's 87.6%). Good pick when cost matters more than benchmark leadership and your workload is pure text. See side-by-side comparisons.