Summary
On August 7, 2026 Cloudflare merged Workers AI and AI Gateway into a single control plane. Developers now call both Cloudflare-hosted models and supported third-party providers through the same env.AI.run() binding and REST API, pay for inference with prepaid AI Gateway credits, and get unified logging, caching, and cost observability across providers.
What changed
Cloudflare connected Workers AI and AI Gateway so the same AI binding and REST endpoints reach both Cloudflare-managed GPUs and external providers, with prepaid AI Gateway credits covering Workers AI inference and frontier-model rate limits raised from 20 to 50 requests per minute per account.
Why it matters
Consolidating model access, billing, and observability into one plane lowers the integration cost of mixing Cloudflare-hosted and third-party models, and positions AI Gateway as a neutral routing and spend-control layer rather than a single-vendor inference service.
Evidence excerpt
You can now use prepaid AI Gateway credits to pay for Workers AI inference.