xAI Grok API Pricing
Elon Musk's AI lab offering Grok models with real-time knowledge and deep X/Twitter integration
| Model | Input (per 1M tokens) | Output (per 1M tokens) | Context Window | Tier |
|---|---|---|---|---|
| Grok 4.6 NEW | $2.00 | $6.00 | 500K | Mid |
| Grok 4.5 | $2.00 | $6.00 | 500K | Mid |
| Grok 4.20 | $1.25 | $2.50 | 1M | Mid |
| Grok 4.3 | $1.25 | $2.50 | 1M | Mid |
| Grok Build 0.1 | $1.00 | $2.00 | 256K | Budget |
Why xAI Grok?
Grok models bring unique capabilities with real-time world knowledge and native X/Twitter integration.
Real-Time Knowledge
Grok models have access to real-time information, making them ideal for tasks that require up-to-date knowledge about current events, trends, and breaking news.
X/Twitter Integration
Native integration with the X platform gives Grok unique access to social media data, conversations, and trending topics — useful for sentiment analysis and social listening.
Competitive Pricing
5 models from $1.00 to $2.00 input per 1M tokens. Grok 4.5 ($2.00/$6.00) is the flagship, Grok 4.20/4.3 ($1.25/$2.50) offer strong mid-tier value, and Grok Build 0.1 ($1.00/$2.00) handles high-volume workloads.
Which Grok Model Should You Use?
Four models for different performance and cost requirements.
Grok 4.5
- 500K context window
- Flagship reasoning model
- Most capable Grok model
- Complex multi-step tasks
- Competitive with GPT-5.5 and Claude Opus
Grok 4.20
- 1M context window
- Latest mid-tier model
- Strong reasoning and analysis
- Real-time world knowledge
- Same price as Grok 4.3, newer architecture
Grok 4.3
- 1M context window
- Proven mid-tier model
- Strong reasoning and analysis
- Real-time world knowledge
- Established track record
Grok Build 0.1
- 256K context window
- Ultra-fast response times
- Classification and extraction
- High-volume workloads
- Budget-friendly pricing
xAI Grok Cost Calculator
Estimate your monthly spend on xAI Grok models.
Monthly Cost Estimate
Based on published API pricing. Actual costs may vary.
Need the full calculator with all providers? Compare xAI against OpenAI, Anthropic, Google, and more.
Try the Full APIpulse CalculatorxAI Grok vs Competitors
How xAI Grok pricing stacks up against other major LLM providers.
| Provider | Model | Input (per 1M tokens) | Output (per 1M tokens) | Context Window |
|---|---|---|---|---|
| xAI | Grok 4.5 | $2.00 | $6.00 | 500K |
| xAI | Grok 4.20 | $1.25 | $2.50 | 1M |
| xAI | Grok 4.3 | $1.25 | $2.50 | 1M |
| OpenAI | GPT-4o | $2.50 | $10.00 | 128K |
| Anthropic | Claude Sonnet 4.6 | $3.00 | $15.00 | 200K |
| DeepSeek | V4 Pro | $0.66 | $1.98 | 1M |
| xAI | Grok Build 0.1 | $1.00 | $2.00 | 256K |
| OpenAI | GPT-4o mini | $0.15 | $0.60 | 128K |
| Gemini 2.5 Flash-Lite | $0.10 | $0.40 | 1M | |
| DeepSeek | V4 Flash | $0.22 | $0.28 | 128K |
See how xAI Grok models stack up against OpenAI, Anthropic, and more
Tool and Batch Costs: What the Token Table Leaves Out
An xAI agent request can have more than one billable component. Model input, cached, reasoning and output tokens are charged at the selected model's rate. xAI-hosted tools can then add an invocation charge, while media generation and storage use their own units.
| Component | Current published rate | Cost rule |
|---|---|---|
| Web Search | $5 / 1,000 calls | In addition to model tokens |
| X Search | $5 / 1,000 calls through Sep 20, 2026 | From Sep 21, 2026: $5/1,000 posts fetched and $10/1,000 user profiles fetched |
| Code Execution | $5 / 1,000 calls | In addition to model tokens |
| Attachment Search | $10 / 1,000 calls | In addition to model tokens |
| Collections Search | $2.50 / 1,000 calls | In addition to model tokens and storage |
| Remote MCP tools | Token-based | No separate xAI invocation fee; remote provider costs may still apply |
The model may choose multiple tool calls, so a single user request does not equal one tool invocation. Image and video understanding for search results is token-billed; image generation follows Imagine API rates.
xAI Batch API pricing
Batch requests are typically completed within 24 hours and do not count against real-time rate limits. The discount is not a blanket rate across every model: xAI currently publishes a 20% discount for Grok 4.3 and the Grok 4.20 variants only, applied to all token types (input, cached, output, reasoning). Models not listed by xAI have no Batch discount, even if they accept Batch requests. Image and video Batch jobs remain at standard media rates. Source: xAI Pricing docs (last updated September 7, 2026).
Related Reading
Deep dives into AI pricing, provider comparisons, and cost optimization.
Grok 4.3 vs GPT-5.4: Which Mid-Tier AI API Wins?
GPT-5.4 costs 6x more on output tokens than Grok 4.3. Real pricing breakdown, use cases, and when to pick each.
Read the comparison →DeepSeek V4 Pricing Breakdown: The Cheapest Frontier Model?
A detailed look at DeepSeek V4 Pro and V4 Flash pricing, and how they compare to GPT-4o and Claude Sonnet.
Read the full breakdown →Best LLM APIs for Startups in 2026
Which AI API providers offer the best value for early-stage startups? We rank every major option by price, quality, and developer experience.
Read the full ranking →How to Switch LLM Providers Without Breaking Your App
A practical guide to migrating between AI API providers while minimizing downtime and code changes.
Read the full guide →Calculate Your xAI Grok Costs
See exactly what xAI Grok will cost for your specific workload. Compare against other providers in seconds.
Go to Calculator →Get notified when xAI prices change
No spam. Only pricing updates and new features. Unsubscribe anytime.