Coding & apps
Cloudflare Workers AI
Run text, image, speech, embedding and Clef decision models on Cloudflare, with 10,000 free Neurons each day.
Start for free
Workers Free
Renews every 1 day (calendar period); Does not roll over
Standard Workers AI billing.
Feature quotas and conditions
Frontier paid models require Workers Paid or prepaid AI Gateway credits.
Requests stop when the daily allowance is exhausted. Upgrade to Workers Paid to continue.
View on official sitePlans and usage
Workers Paid
Renews every 1 day (calendar period); Does not roll over
Standard Workers AI billing.
Feature quotas and conditions
Minimum $5/account/month; usage of other Cloudflare services is billed separately.
$0.011/1,000 Neurons above the daily allowance.
View on official siteAI Gateway prepaid
Feature quotas and conditions
Top up manually or configure automatic top-ups.
Load credits, set Workers AI Billing to Unified billing and include the gateway in the request.
Negative balances are charged to the saved payment method at the beginning of the next month.
View on official siteWorkers Paid starts at $5/account/month. Inference above the free daily allowance costs $0.011/1,000 Neurons. AI Gateway prepaid credit purchases have a 5% fee.
Additional usage rates
| Available plans | Quantity | Price | Terms |
|---|---|---|---|
| AI Gateway prepaid / Workers Paid | 1,000 characters | $0.015 / 1,000 characters | aura-1: input characters |
| AI Gateway prepaid / Workers Paid | 1,000 characters | $0.03 / 1,000 characters | aura-2-en: input characters |
| AI Gateway prepaid / Workers Paid | 1,000 characters | $0.03 / 1,000 characters | aura-2-es: input characters |
| AI Gateway prepaid / Workers Paid | 10 generation steps | $0.0011 / 10 generation steps | flux-1-schnell: generation steps |
| AI Gateway prepaid / Workers Paid | 10 512×512 tiles | $0.0005 / 10 512×512 tiles | flux-1-schnell: 512×512 tiles |
| AI Gateway prepaid / Workers Paid | 1 input tile-steps | $0.0002 / input tile-steps | flux-2-dev: input tile-steps |
| AI Gateway prepaid / Workers Paid | 1 output tile-steps | $0.0004 / output tile-steps | flux-2-dev: output tile-steps |
| AI Gateway prepaid / Workers Paid | 1 512×512 tiles | $0.0001 / 512×512 tiles | flux-2-klein-4b: input 512×512 tiles |
| AI Gateway prepaid / Workers Paid | 1 512×512 tiles | $0.0003 / 512×512 tiles | flux-2-klein-4b: output 512×512 tiles |
| AI Gateway prepaid / Workers Paid | 1 first output megapixel | $0.015 / first output megapixel | flux-2-klein-9b: first output megapixel |
| AI Gateway prepaid / Workers Paid | 1 input image megapixels | $0.002 / input image megapixels | flux-2-klein-9b: input image megapixels |
| AI Gateway prepaid / Workers Paid | 1 subsequent output megapixels | $0.002 / subsequent output megapixels | flux-2-klein-9b: subsequent output megapixels |
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0077 / 60 seconds | flux: input audio seconds |
| AI Gateway prepaid | 1 prepaid USD credits | $1.00 / prepaid USD credits | A 5% purchase fee is added to each dollar of prepaid credits. Gateway uses Unified billing and has prepaid AI Gateway credits. |
| 1 inference requests | Free | Fine-tuned inference using an adapter on a supported LoRA model during open beta. | |
| AI Gateway prepaid / Workers Paid | 1 generation steps | $0.0001 / generation steps | lucid-origin: generation steps |
| AI Gateway prepaid / Workers Paid | 1 512×512 tiles | $0.007 / 512×512 tiles | lucid-origin: 512×512 tiles |
| 1 converted files | Free | Supported non-image document formats convert for free. | |
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0002 / 60 seconds | melotts: output audio seconds |
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0052 / 60 seconds | nova-3: input audio seconds |
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0092 / 60 seconds | nova-3: input audio seconds |
| AI Gateway prepaid / Workers Paid | 1 generation steps | $0.0001 / generation steps | phoenix-1.0: generation steps |
| AI Gateway prepaid / Workers Paid | 1 512×512 tiles | $0.0058 / 512×512 tiles | phoenix-1.0: 512×512 tiles |
| AI Gateway prepaid / Workers Paid | 1,000,000 image inferences | $2.51 / 1,000,000 image inferences | resnet-50: image inferences |
| AI Gateway prepaid / Workers Paid | 6,000 seconds | $0.0338 / 6,000 seconds | smart-turn-v2: input audio seconds |
| 1 generation steps | Free | ||
| 1 generation steps | Free | ||
| 1 generation steps | Free | ||
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0005 / 60 seconds | whisper: input audio seconds |
| AI Gateway prepaid / Workers Paid | 60 seconds | $0.0005 / 60 seconds | whisper-large-v3-turbo: input audio seconds |
| Workers Paid | 1,000 Neurons | $0.011 / 1,000 Neurons |
Choose your workflow
Developers adding chat, vision, image generation, transcription, speech synthesis, embeddings and content classification to applications.
Create a Cloudflare account and choose a Workers AI model. Add an AI binding and call env.AI.run from a Worker, or create a Workers AI API token and call the REST API with your account ID. For the OpenAI SDK, set baseURL to https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/v1.
Account eligibility follows Cloudflare export control and sanctions terms.
Workers Free
Asynchronous batch inference
IncludedLoRA fine-tuned inference
IncludedUpload trained adapters for inference; free during open beta.
Convert documents to Markdown or plain text
IncludedMost formats convert for free; image conversion model inference follows Workers AI usage pricing.
Clef typed decisions and probabilities
IncludedText embeddings and reranking
IncludedImage generation
IncludedSpeech recognition and generation
IncludedText generation and vision
IncludedWorkers Paid
Asynchronous batch inference
IncludedLoRA fine-tuned inference
IncludedUpload trained adapters for inference; free during open beta.
Convert documents to Markdown or plain text
IncludedMost formats convert for free; image conversion model inference follows Workers AI usage pricing.
Clef typed decisions and probabilities
IncludedText embeddings and reranking
IncludedImage generation
IncludedSpeech recognition and generation
IncludedText generation and vision
IncludedAI Gateway prepaid
Asynchronous batch inference
IncludedLoRA fine-tuned inference
IncludedUpload trained adapters for inference; free during open beta.
Convert documents to Markdown or plain text
IncludedMost formats convert for free; image conversion model inference follows Workers AI usage pricing.
More capabilities and limits
Data retention · You own inputs, outputs and embeddings. They are not used for training without explicit consent. R2, KV, Durable Objects or Vectorize can store content when configured.
Data retention · AI Gateway logs requests and responses by default. Disable whole logs or retain metadata only. Accounts creating their first gateway from September 24, 2026 use Workers Logs.
Workers AI requests routed through AI Gateway; logging follows gateway settings or request headers.
Payment processing fee · 5 %
Purchases of prepaid AI Gateway credits in USD.
Models and access
indictrans2-en-indic-1B
Gateway uses Unified billing and has prepaid AI Gateway credits.
gemma-sea-lion-v4-27b-it
Gateway uses Unified billing and has prepaid AI Gateway credits.
bge-base-en-v1.5
Gateway uses Unified billing and has prepaid AI Gateway credits.
bge-large-en-v1.5
Gateway uses Unified billing and has prepaid AI Gateway credits.
bge-m3
Gateway uses Unified billing and has prepaid AI Gateway credits.
bge-reranker-base
Gateway uses Unified billing and has prepaid AI Gateway credits.
bge-small-en-v1.5
Gateway uses Unified billing and has prepaid AI Gateway credits.
flux-1-schnell
Gateway uses Unified billing and has prepaid AI Gateway credits.
flux-2-dev
Gateway uses Unified billing and has prepaid AI Gateway credits.
flux-2-klein-4b
Gateway uses Unified billing and has prepaid AI Gateway credits.
flux-2-klein-9b
Gateway uses Unified billing and has prepaid AI Gateway credits.
stable-diffusion-xl-lightning
Gateway uses Unified billing and has prepaid AI Gateway credits.
clef-flash
Gateway uses Unified billing and has prepaid AI Gateway credits.
clef-omni
Gateway uses Unified billing and has prepaid AI Gateway credits.
clef
Gateway uses Unified billing and has prepaid AI Gateway credits.
phonellm-mini-alpha-2
Gateway uses Unified billing and has prepaid AI Gateway credits.
aura-1
Gateway uses Unified billing and has prepaid AI Gateway credits.
aura-2-en
Gateway uses Unified billing and has prepaid AI Gateway credits.
aura-2-es
Gateway uses Unified billing and has prepaid AI Gateway credits.
flux
Gateway uses Unified billing and has prepaid AI Gateway credits.
nova-3
Gateway uses Unified billing and has prepaid AI Gateway credits.
deepseek-r1-distill-qwen-32b
Gateway uses Unified billing and has prepaid AI Gateway credits.
deepseek-v4-flash-0731
Gateway uses Unified billing and has prepaid AI Gateway credits.
deepseek-v4-pro-0813
Gateway uses Unified billing and has prepaid AI Gateway credits.
embeddinggemma-300m
Gateway uses Unified billing and has prepaid AI Gateway credits.
gemma-2b-it-lora
Gateway uses Unified billing and has prepaid AI Gateway credits.
gemma-4-26b-a4b-it
Gateway uses Unified billing and has prepaid AI Gateway credits.
gemma-7b-it-lora
Gateway uses Unified billing and has prepaid AI Gateway credits.
distilbert-sst-2-int8
Gateway uses Unified billing and has prepaid AI Gateway credits.
granite-4.0-h-micro
Gateway uses Unified billing and has prepaid AI Gateway credits.
lucid-origin
Gateway uses Unified billing and has prepaid AI Gateway credits.
phoenix-1.0
Gateway uses Unified billing and has prepaid AI Gateway credits.
llava-1.5-7b-hf
Gateway uses Unified billing and has prepaid AI Gateway credits.
dreamshaper-8-lcm
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-2-7b-chat-hf-lora
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-3.1-8b-instruct-fp8
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-3.2-11b-vision-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-3.2-1b-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-3.2-3b-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-3.3-70b-instruct-fp8-fast
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-4-scout-17b-16e-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
llama-guard-3-8b
Gateway uses Unified billing and has prepaid AI Gateway credits.
m2m100-1.2b
Gateway uses Unified billing and has prepaid AI Gateway credits.
resnet-50
Gateway uses Unified billing and has prepaid AI Gateway credits.
mistral-7b-instruct-v0.2-lora
Gateway uses Unified billing and has prepaid AI Gateway credits.
mistral-small-3.1-24b-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
moondream3.1-9B-A2B
Gateway uses Unified billing and has prepaid AI Gateway credits.
kimi-k2.6
Gateway uses Unified billing and has prepaid AI Gateway credits.
kimi-k2.7-code
Gateway uses Unified billing and has prepaid AI Gateway credits.
melotts
Gateway uses Unified billing and has prepaid AI Gateway credits.
nemotron-3-120b-a12b
Gateway uses Unified billing and has prepaid AI Gateway credits.
gpt-oss-120b
Gateway uses Unified billing and has prepaid AI Gateway credits.
gpt-oss-20b
Gateway uses Unified billing and has prepaid AI Gateway credits.
whisper-large-v3-turbo
Gateway uses Unified billing and has prepaid AI Gateway credits.
whisper-tiny-en
Gateway uses Unified billing and has prepaid AI Gateway credits.
whisper
Gateway uses Unified billing and has prepaid AI Gateway credits.
plamo-embedding-1b
Gateway uses Unified billing and has prepaid AI Gateway credits.
smart-turn-v2
Gateway uses Unified billing and has prepaid AI Gateway credits.
qwen2.5-coder-32b-instruct
Gateway uses Unified billing and has prepaid AI Gateway credits.
qwen3-30b-a3b-fp8
Gateway uses Unified billing and has prepaid AI Gateway credits.
qwen3-embedding-0.6b
Gateway uses Unified billing and has prepaid AI Gateway credits.
qwen3.8-27b
Gateway uses Unified billing and has prepaid AI Gateway credits.
qwq-32b
Gateway uses Unified billing and has prepaid AI Gateway credits.
stable-diffusion-v1-5-inpainting
Gateway uses Unified billing and has prepaid AI Gateway credits.
stable-diffusion-xl-base-1.0
Gateway uses Unified billing and has prepaid AI Gateway credits.
apertus-v1.5-8b
Complete the official access request form to use this model.
Complete the official access request form to use this model.
Complete the official access request form to use this model.
eurollm-9b-it
Complete the official access request form to use this model.
Complete the official access request form to use this model.
Complete the official access request form to use this model.
glm-4.7-flash
Gateway uses Unified billing and has prepaid AI Gateway credits.
glm-5.2
Gateway uses Unified billing and has prepaid AI Gateway credits.
glm-5.3-flash
Gateway uses Unified billing and has prepaid AI Gateway credits.
glm-5.3
Gateway uses Unified billing and has prepaid AI Gateway credits.
gemma-3-12b-it
llama-2-7b-chat-fp16
llama-3-8b-instruct-awq
llama-3-8b-instruct
llama-3.1-70b-instruct-fp8-fast
llama-3.1-8b-instruct-awq
llama-3.1-8b-instruct-fp8-fast
llama-3.1-8b-instruct
mistral-7b-instruct-v0.1
kimi-k2.5
Model rates
| Model / tier | Input | Cache read | Cache write | Output |
|---|---|---|---|---|
| indictrans2-en-indic-1B | 0.342 | — | — | 0.342 |
| gemma-sea-lion-v4-27b-it | 0.351 | — | — | 0.555 |
| bge-base-en-v1.5 | 0.0666 | — | — | — |
| bge-large-en-v1.5 | 0.204 | — | — | — |
| bge-m3 | 0.0118 | — | — | — |
| bge-reranker-base | 0.00311 | — | — | — |
| bge-small-en-v1.5 | 0.0202 | — | — | — |
| clef-flash | 0.038 | — | — | 0 |
| clef-omni | 0.15 | — | — | 0 |
| clef | 0.24 | — | — | 0 |
| phonellm-mini-alpha-2 | 0.06 | — | — | 0.24 |
| deepseek-r1-distill-qwen-32b | 0.497 | — | — | 4.881 |
| deepseek-v4-flash-0731 | 0.44 | 0.014 | — | 1.32 |
| deepseek-v4-pro-0813 | 1.32 | 0.044 | — | 3.96 |
| gemma-3-12b-it | 0.345 | — | — | 0.556 |
| gemma-4-26b-a4b-it | 0.1 | — | — | 0.3 |
| distilbert-sst-2-int8 | 0.0263 | — | — | — |
| granite-4.0-h-micro | 0.017 | — | — | 0.112 |
| llama-2-7b-chat-fp16 | 0.556 | — | — | 6.667 |
| llama-3-8b-instruct-awq | 0.123 | — | — | 0.266 |
| llama-3-8b-instruct | 0.282 | — | — | 0.827 |
| llama-3.1-70b-instruct-fp8-fast | 0.293 | — | — | 2.253 |
| llama-3.1-8b-instruct-awq | 0.123 | — | — | 0.266 |
| llama-3.1-8b-instruct-fp8-fast | 0.045 | — | — | 0.384 |
| llama-3.1-8b-instruct-fp8 | 0.152 | — | — | 0.287 |
| llama-3.1-8b-instruct | 0.282 | — | — | 0.827 |
| llama-3.2-11b-vision-instruct | 0.0485 | — | — | 0.676 |
| llama-3.2-1b-instruct | 0.027 | — | — | 0.201 |
| llama-3.2-3b-instruct | 0.0509 | — | — | 0.335 |
| llama-3.3-70b-instruct-fp8-fast | 0.293 | — | — | 2.253 |
| llama-4-scout-17b-16e-instruct | 0.27 | — | — | 0.85 |
| llama-guard-3-8b | 0.484 | — | — | 0.03 |
| m2m100-1.2b | 0.342 | — | — | 0.342 |
| mistral-7b-instruct-v0.1 | 0.11 | — | — | 0.19 |
| mistral-small-3.1-24b-instruct | 0.351 | — | — | 0.555 |
| moondream3.1-9B-A2B | 0.3 | — | — | 1 |
| kimi-k2.5 | 0.6 | 0.1 | — | 3 |
| kimi-k2.6 | 0.95 | 0.16 | — | 4 |
| kimi-k2.7-code | 0.95 | 0.19 | — | 4 |
| nemotron-3-120b-a12b | 0.5 | — | — | 1.5 |
| gpt-oss-120b | 0.35 | — | — | 0.75 |
| gpt-oss-20b | 0.2 | — | — | 0.3 |
| plamo-embedding-1b | 0.0186 | — | — | — |
| qwen2.5-coder-32b-instruct | 0.66 | — | — | 1 |
| qwen3-30b-a3b-fp8 | 0.0509 | — | — | 0.335 |
| qwen3-embedding-0.6b | 0.0118 | — | — | — |
| qwen3.8-27b | 0.45 | 0.05 | — | 3.2 |
| qwq-32b | 0.66 | — | — | 1 |
| glm-4.7-flash | 0.0605 | — | — | 0.4 |
| glm-5.2 | 1.4 | 0.26 | — | 4.4 |
| glm-5.3-flash | 0.15 | 0.03 | — | 0.5 |
| glm-5.3 | 1.4 | 0.26 | — | 4.4 |
Official sources
2026-10-16