News Cloudflare
No Cloudflare Workers AI Inference Pricing Update Confirmed for December 2026
Cloudflare documentation shows no announced change to Workers AI inference pricing effective December 2026.
Cloudflare's official documentation as of November 17, 2026, does not contain any announcement of a general inference pricing update for Workers AI taking effect in December 2026.
Current Pricing Structure
Workers AI continues to meter usage through a Neuron-based system at $0.011 per 1,000 Neurons. Every account receives a daily free allocation of 10,000 Neurons that resets at 00:00 UTC. Usage beyond this allocation on the Workers Paid plan is billed at the same rate, while free-plan accounts see requests fail once the allocation is exhausted.

Per-model token pricing remains published with equivalent Neuron values for transparency. Examples include Llama 3.2 1B at $0.027 input and $0.201 output per million tokens, Llama 3.2 3B at $0.051/$0.335, and larger models such as Gemma 4 26B at $0.10/$0.30. Some frontier models require a Workers Paid plan or prepaid AI Gateway credits.
Recent and Upcoming Changes
Model deprecations scheduled for November 20, 2026, affect several older DeepSeek, Qwen, Kimi, and Nemotron variants, while new models including Inkling 2 Small and EmbeddingGemma 2 have been added. AI Search billing, a separate product, has been postponed to December 1, 2026.
Billing shifted earlier from pure Neuron metering to published per-model token equivalents to simplify comparisons. The service remains serverless with no idle GPU charges. Users monitoring spend can expect continued catalog refreshes and possible additions of limits or alerts in future updates.
Frequently asked questions
Has Cloudflare announced new Workers AI pricing for December 2026?
No verified announcement exists in current documentation.
What is the current daily free allocation for Workers AI?
Every account receives 10,000 Neurons per day that reset at 00:00 UTC.
How are Workers AI costs calculated across models?
Costs use published per-model token rates with equivalent Neuron values shown for each model.
Which models require a paid plan or prepaid credits?
Some frontier models require the Workers Paid plan or AI Gateway prepaid credits.
Comments 1