HeadsUpAI

Cloudflare adds xAI Grok models to AI Gateway with unified billing

Cloudflare has integrated the xAI Grok model family into its AI Gateway. This update provides access to 10 models, including the agentic grok-4.3 and the high-context grok-4.20 series. The integration spans text, speech, and media, all accessible through a single interface without requiring separate xAI API keys or complex environment setups.
Models
10 Grok variants (text, speech, image, video)
Top Context Window
2M tokens (grok-4.20 multi-agent)
Billing
Direct through Cloudflare
Authentication
Keyless — no separate xAI API keys
Capabilities
Function calling, structured outputs, reasoning effort

This partnership positions Cloudflare as a central control plane for frontier AI, following its recent addition of GPT-5.5 and MiniMax M3. By hosting grok-4.20-multi-agent-0309 with its 2M-token context window (the data a model processes at once), Cloudflare competes with Vercel to provide infrastructure for complex agentic workflows.

You can now deploy grok-4.20-0309-reasoning for tasks requiring extended thinking (internal deliberation before answering) or grok-imagine-video-1.5-preview for media generation. All usage is billed directly through Cloudflare. Models support function calling, structured outputs, and configurable reasoning effort to balance speed and accuracy.

Cloudflare Developers
Cloudflare Developers
@CloudflareDev
X

We're partnering with @xai to bring Grok to @Cloudflare AI Gateway. • Grok LLMs, audio, image, and video models are now available through AI Gateway • Billed directly through Cloudflare • No additional auth, env, API keys needed https://t.co/DnqsvKDLon

131retweets1.1klikes
View on X

Still wondering? A few quick answers below.

Cloudflare has added xAI's Grok model family to its AI Gateway, a platform that acts as a single control plane for AI applications. This integration allows developers to route requests to Grok models for text, image, and video generation while benefiting from Cloudflare's unified logging, caching, and security features.

The integration includes 10 models from xAI, including the agentic grok-4.3 and the high-context grok-4.20 series. Developers can choose between reasoning variants for complex logic, non-reasoning variants for speed, and specialized models for generative media, such as grok-imagine-video and grok-tts for high-fidelity speech synthesis.

One of the primary benefits of this partnership is unified billing. Instead of managing a separate account and payment method with xAI, all inference costs are billed directly through your existing Cloudflare account. This simplifies financial management for teams using multiple AI providers through the AI Gateway platform.

No. The integration is designed to be keyless for the developer. Once configured within the Cloudflare AI Gateway, you do not need to manage additional xAI authentication, environment variables, or API keys. Cloudflare handles the underlying connection to xAI's infrastructure, streamlining the deployment process for production applications.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Share this update