Grok Models Now Available on Cloudflare's AI Gateway
xAI announced on June 3, 2026 that its Grok models are now available through Cloudflare's AI Gateway. Developers can access Grok's LLM, audio, image, and video generation models from a single Cloudflare dashboard.1
On the same day, Cloudflare Developers (@CloudflareDev) posted "We're partnering with @xai to bring Grok to @Cloudflare AI Gateway," adding details about the partnership. The integration supports OpenAI-compatible endpoints, and can be used simply by replacing the existing xAI API (https://api.x.ai/v1) with the Gateway URL. Specifically, an endpoint such as https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/grok is provided, calling models by specifying names like grok-4. A Workers AI-style call such as env.AI.run("xai/grok-imagine-video-1.5-preview") is also shown as an example.23
Cloudflare AI Gateway is an AI inference proxy service whose beta was announced in September 2023, providing unified access to multiple LLM providers along with caching, rate limiting, observability, and fallback. An August 2025 refresh strengthened direct connectivity and unified billing across six providers—including Anthropic, Google, Groq, OpenAI, and xAI—and more than 350 models. With this xAI partnership, not only Grok's language models but also the Grok Imagine image generator, the grok-imagine-video-1.5-preview video generator, and TTS/STT audio models can now be handled via the Gateway.45
Supported models are wide-ranging, including grok-4.3 with a 1M-token context window and the multi-agent grok-4.20 series for text generation, Grok Imagine at a default 2k resolution for images, audio-synchronized models for video, and grok-tts capable of laughter, whispers, and pauses for audio. Billing is consolidated through the Cloudflare account, eliminating the need for separate per-provider billing. Features such as virtual key management and dynamic routing that optimizes cost, performance, and user experience are also included. The Gateway's core features are available even on the free plan, with paid tiers for scaling up.6
Competitors include similar services such as Vercel AI Gateway and Portkey, but Cloudflare's strengths lie in low latency and caching leveraging its global edge network, and seamless integration with Workers AI. In response to the announcement, the Grok official account reacted with "This is awesome! Thrilled Grok models (LLMs, audio, image + video gen) are now live... Devs are about to cook." As the announcement is fresh, detailed use-case feedback remains limited, but sentiment centers on appreciation for seamless access without additional keys and the convenience of unified billing, with no posts about limitations or stumbling blocks observed so far.7