Cloud & Architecture · Freemium
Cloudflare Workers AI
Cloudflare Workers AI is Cloudflare's serverless inference platform for running open and partner AI models on GPUs across Cloudflare's global network. Developers call models from a Worker through the AI binding (env.AI.run) or from any backend through a REST API, without renting, provisioning or scaling GPUs themselves. Model IDs follow the pattern @cf/<provider>/<model>, for example @cf/openai/gpt-oss-120b or @cf/zai-org/glm-5.3.
Who it is for: teams already building on Cloudflare Workers, Pages, Durable Objects or the Agents SDK who want inference close to their application code, and developers who want pay-per-use access to open-weight models behind a single account.
Key features: - A catalog that covers text generation (LLMs from DeepSeek, Z.ai GLM, Qwen, OpenAI gpt-oss, Google Gemma, Meta Llama, Mistral, IBM Granite and others), embeddings, reranking, speech-to-text, text-to-speech, image generation, translation and classification. - Workers binding and REST API access, with OpenAI-compatible endpoints for many text models. - Function calling, structured outputs and long context on supported models, plus LoRA adapter inference on dedicated base models. - Integration with Cloudflare AI Gateway for caching, rate limiting, logging, fallbacks and model aliases, and with Vectorize and AI Search for retrieval.
Pricing type: usage-based (freemium). Workers AI is included on both Workers Free and Workers Paid plans with a free allocation of 10,000 Neurons per day; usage above that requires Workers Paid and is billed at $0.011 per 1,000 Neurons, with per-model token prices published on the pricing page. Some frontier models require a paid billing method or prepaid AI Gateway credits.
Model lifecycle: Cloudflare periodically refreshes the catalog and deprecates older models. Deprecated models stop responding after their retirement date and return an HTTP 410 error, so production apps should keep model IDs in configuration and watch the Workers AI changelog.
Alternatives: Together AI, Fireworks AI, Groq, Amazon Bedrock, Google Vertex AI, OpenRouter (as a router across hosts), and Vercel AI Gateway for multi-provider access.
Official site: https://developers.cloudflare.com/workers-ai/

Cloud · Comparison · Metarix Editorial · Nov 11, 2026 · 12 min
Serverless GPU Inference Compared: Pricing and Cold Starts
Modal, Runpod Serverless, Replicate, Baseten and Cloud Run with GPUs compared on list price, cold starts and idle billing, with a worked cost example and a decision flow for picking one.
View comparison →
Cloud · Analysis · Metarix Editorial · Nov 9, 2026 · 9 min
Workers AI Retires 16 LLMs on Nov 20: Migration Guide
Cloudflare will switch off 16 Workers AI LLMs, including Kimi K2.6, GLM-5.2 and both DeepSeek V4 snapshots, on November 20, 2026. Here are the replacements, the cost impact and an 11-day migration plan.
Read analysis →