Cloudflare Developer Platform Pricing & Limits
Unofficial list compiled from information on developers.cloudflare.com about each of the product
WorkersCron TriggersDurable ObjectsPagesWorkers KVR2D1QueuesHyperdriveWorkers Analytics EngineWorkers LogpushTail WorkersWorkers AIVectorizeAI GatewayImagesStreamCalls
Workers Free
Workers Paid
Compute
Requests
100,000 / day
10,000,000 / month+$0.30 / 1,000,000 requests
CPU time
-
30,000,000 ms / month+$0.02 / 1,000,000 ms
CPU time limit
10 ms / invocation
30 s / invocation
Subrequests
This includes requests to internet using `fetch`, requests to Cloudflare services like R2, KV, D1 or requests to Cache API
50 / request
1,000 / request
Simultaneous outgoing requests/ connections
6
Environment variables
64 / Worker
128 / Worker
Environment variable size
5 KB
Worker size
1 MB
10 MB
Worker startup time
400 ms
Workers per account
100
500
Worker memory
128 MB
Cache API maximum object size
512 MB
Cache API calls per request
This limits are the same and count towards Subrequests limit
50
1000
Total size of the logs
This includes size of `console.log` statements, exceptions, request metadata and headers
128 KB
Durable Objects
Only on Workers PaidRequests
This includes HTTP requests, RPC sessions, WebSocket messages, and alarm invocations
-
1,000,000 / month+$0.15 / 1,000,000 requests
Duration
-
400,000 GB-s+$12.50 / 1,000,000 GB-s
Memory size allocation
Each Durable Object allocate 128 MB of memory to run regardless of actual usage
-
128 MB
Durable Object namespaces per account
-
500 (identical to Workers limit)
Storage per account
-
50 GB (can be raised)
Key size
-
2 KiB (2,048 bytes)
Value size
-
128 KiB (131,072 bytes)
WebSocket incoming message size
-
1 MiB
WebSocket incoming messages
Incoming WebSocket protocol pings are not billed
-
20 messages billed as 1 request
WebSocket outgoing messages
-
Unlimited
CPU time per request
Each incoming HTTP request or WebSocket message resets the remaining available CPU time to 30 seconds
-
30 s
Transactional Storage API Reads
-
1,000,000 request units+$0.20 / 1,000,000 request units
Transactional Storage API Writes
-
1,000,000 request units+$1 / 1,000,000 request units
Transactional Storage API Deletes
-
1,000,000 requests+$1 / 1,000,000 requests
Transactional Storage API Stored data
-
1 GB+$0.20 / 1 GB-month
Request unit size
As example, reading or writing 9 KB of data in 1 request will consume 3 request units
-
4 KB
Static assets requests
Unlimited
Pages Functions requests
Billed as Workers requests
Builds per account
500 / month on Free plan5,000 / month on Pro plan20,000 / month on Business plan
Concurrent builds per account
1 on Free plan5 on Pro plan20 on Business plan
Custom domains per project
100 on Free plan250 on Pro plan500 on Business plan
Build timeout
20 min
Preview deployments
Unlimited
Pages projects per account
Soft limit of 100 projects
Files per project
20,000
Maximum file size
25 MiB
Headers
Custom headers to attach to static assets
100 header rules
Redirects
Custom redirects for static projects
2,000 static redirects100 dynamic redirects
Storage
Read request
100,000 / day
10,000,000 / month+$0.50 / 1,000,000 requests
Write request
1,000 / day
1,000,000 / month+$5 / 1,000,000 requests
Delete request
1,000 / day
1,000,000 / month+$5 / 1,000,000 requests
List request
1,000 / day
1,000,000 / month+$5 / 1,000,000 requests
Stored data
1 GB
1 GB+$0.50 / 1 GB-month
Writes to the same key
1 / second
Operations per Worker invocation
Must be wrong, because this regulated by Workers Subrequests limit
1000
KV namespaces per account
200
Key size
512 bytes
Key metadata size
1024 bytes
Value size
25 MiB
Standard storage
10 GB / month
10 GB / month+$0.015 / 1 GB-month
Standard storage Class A Operations
Operation to mutate data, such as writes
1,000,000 requests / month
1,000,000 requests / month+$4.50 / 1,000,000 requests
Standard storage Class B Operations
Opeations to read existing data
10,000,000 requests / month
10,000,000 requests / month+$0.36 / 1,000,000 requests
Infrequent Access storage
-
$0.010 / 1 GB-month
Infrequent Access storage Class A Operations
Operation to mutate data, such as writes
-
$9 / 1,000,000 requests
Infrequent Access storage Class B Operations
Opeations to read existing data
-
$0.90 / 1,000,000 requests
Infrequent Access storage Data retrieval fee
-
$0.01 / 1 GB
Infrequent Access storage Minimum storage duration
-
30 days
Ergees
Free
R2 buckets per account
1,000
Object key size
1,024 bytes
Object metadata size
8,192 bytes
Object size
4,995 TiB
Maximum upload size in 1 request
4.995 GiB
Maximum upload parts
10,000
Custom domains per bucket
50
Rows reads
5,000,000 / day
25 billion / month+$0.001 / 1,000,000 rows
Rows writes
100,000 / day
50,000,000 / month+$1 / 1,000,000 rows
Storage per account
5 GB
5 GB+$0.75 / 1 GB-month
D1 databases per account
10
50,000
Maximum database size
500 MB
10 GB
Maximum storage per account
5 GB
250 GB (can be increased)
Time Travel duration
7 days
30 days
Maximum Time Travel restore operations
10 restores per 10 minute per database
Maximum number of columns per table
100
Maximum table row size
1 MB
Maximum SQL statement length
100 KB
Maximum bound parameters per query
100
Maximum arguments per SQL function
32
Maximum bytes in a LIKE or GLOB pattern
50 bytes
Maximum bindings per Worker script
Approximately 5,000
Maximum SQL query duration
30 s
Standard operations
-
1,000,000 operations / month+$0.40 / 1,000,000 operations
Operation size
As example, read, write or delete action on 65 KB message will counts as 2 operations
-
64 KB
Queues per account
-
10,000 (while in beta)
Message size
-
128 KB (including up to 100 bytes of internal metadata)
Message retries
-
100
Maximum consumer batch size
-
100 messages (while in beta)
Maximum messages per `sendBatch` call
-
100 (or 256KB in total)
Batch wait time
-
60 s
Per-queue message throughput
-
400 messages / second (while in beta)
Message retention period
-
4 days (96 hours)
Per-queue backlog size
-
25 GB
Concurrent consumer invocations
-
20 (push-based only)
Duration of consumer invocation
-
15 min
`visibilityTimeout` (pull-based queues)
-
12 hours
`delaySeconds` (when sending or retrying)
-
12 hours
Requests to the Queues API
-
1,200 requests / 5 min
Hyperdrive
Only on Workers PaidUsage of Hyperdrive
-
Free
Maximum configured databases
-
25 / account
Initial connection timeout
-
15 s
Idle connection timeout
-
10 min
Maximum cached query response size
-
50 MB
Maximum query duration
-
60 s
Maximum username & database name length
-
63 bytes
Maximum origin database connections per region
-
10 - 20
Maximum potential origin database connections
-
Approximately 80 - 100
Workers Analytics Engine
Free while in BetaData points written
100,000 / day
10,000,000 / month+$0.25 / 1,000,000 data points
Read queries
10,000 / day
1,000,000 / month+$1 / 1,000,000 queries
Data retention
3 month
Payload limit per 1 data point
12 blobs, 12 doubles and 1 index
Maximum size of all blobs in data point
5,120 bytes
Maximum index size
96 bytes
Maximum data points written
25 / Worker invocation
Observability
Workers Logpush
Only on Workers PaidRequests
-
10,000,000 / month+$0.05 / 1,000,000 requests
Message size
-
2,056 / log line
Array limit
-
20 elements
Log message array
-
A nested array with a limit of 20 elements
Tail Workers
Only on Workers PaidCPU time
-
Billed as regular Workers
Requests
-
Not billed for requests
AI
Usage of non-beta models on Workers AI
10,000 Neurons / day
10,000 Neurons / day+$0.011 / 1,000 Neurons
Usage of beta models on Workers AI
Unlimited
Price example of Speech Recognition
-
whisper: $0.0022 / 1 min of audio
Price example of Image Classification
-
Resnet-50: $0.0000025 / 1 image
Price example of Text Classification
-
distilbert-sst-2-int8: $0.33 / 1M input tokens
Price example of Text Embeddings
-
bge-small-en-v1.5: $0.003 / 1M input tokensbge-base-en-v1.5: $0.014 / 1M input tokensbge-large-en-v1.5: $0.022 / 1M input tokens
Price example of Text Generation
-
llama-2-7b-chat-fp16: $0.56 / 1M input tokens, $6.66 / 1M output tokensllama-2-7b-chat-int8: $0.16 / 1M input tokens, $0.24 / 1M output tokensmistral-7b-instruct: $0.11 / 1M input tokens, $0.19 / 1M output tokens
Price example of Translation
-
m2m100-1.2b: $0.13 / 1M input tokens, $0.70 / 1M output tokens
Stored vector dimensions
5,000,000
10,000,000+$0.040 / 1,000,000 stored dimensions
Queried vector dimensions
30,000,000 / month
50,000,000 / month+$0.040 / 1,000,000 queried dimensions
Indexes per account
100 (while in beta)
Maximum dimensions per vector
1536 (while in beta)
Maximum vector ID length
64 bytes
Metadata per vector
10 KiB (while in beta)
Maximum returned results (`topK`)
20 (while in beta)
Maximum upsert batch size
1,000 from Workers / 5,000 from HTTP API (while in beta)
Maximum index & namespace names length
63 bytes
Maximum vectors per index
200,000 (while in beta)
Maximum namespaces per index
1,000 (while in beta)
Usage of AI Gateway
Free
Media
Images stored
This is price for images stored within Cloudflare Images. This is not required to use Images product
$5 / 100,000 original images (pre-paid plan)
Images delivered
This is price for delivered images that stored within Cloudflare Images. This is not required to use Images product
$1 / 100,000 images stored in Images
Images transformed
Unique transformation is a request to transform an original image with a combination of flags, exluding `format` flag
$0.50 / 1,000 unique transformations / month
Maximum image size
100 megapixelsGIF/WebP animations up to 50 MP total (sum of all frames)
Maximum file size
100 MB
Generate AVIF format images longest side limit
1,200 pixels & 1,600 when `format=avif` explicitly used640 pixels when system is overloaded
Generate WebP format images longest side limit
2,560 pixels for lossy & 1,920 for lossless when system is overloaded
Generate other format images longest side limit
12,000 pixels
Video stored
$5 / 1,000 min / month
Video delivered
$1 / 1,000 min / month
Video upload size
30 GB
Concurrent uploading videos to Stream
120 (can be raised)