Skip to content
API Pricing

10,634 prices on 4,757 listings · last collected 7 min ago · 399 flagged changes in 30 days

Every price, as each platform publishes it

A collector reads every public price on eight platforms, daily by default, and keeps the page it read. Nothing is typed in by hand. Search a model to see who charges least, per unit, right now.

Showing 1–33 of 33

“Cheapest” is the lowest exact price in this unit for your search. Rows can differ in resolution, length or quality — read the label, or add it to the search (e.g. “1080”).

Per million input tokens14

PlatformModel · what it coversUSDPlatform’s own wordingSince
ReplicatePlatform-hostedibm-granite/granite-3.0-2b-instruct$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-3.1-2b-instruct$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-3.1-8b-instruct$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-3.2-8b-instruct$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-3.3-8b-instruct$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-speech-4.1-2b$0.03/M inCheapest$0.03 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-3.0-8b-instruct$0.05/M in$0.05 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-8b-code-instruct-128k$0.05/M in$0.05 per million input tokens
OpenRouterThird-partyIBM: Granite 4.2 8Bvia CoreWeave · bf16$0.06/M inprompt: 0.00000006 / token (endpoint coreweave/bf16)
ReplicatePlatform-hostedibm-granite/granite-4.0-h-small$0.06/M in$0.06 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-4.1-8b$0.06/M in$0.06 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-4.2-8b$0.06/M in$0.06 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-vision-4.1-4b$0.06/M in$0.06 per million input tokens
ReplicatePlatform-hostedibm-granite/granite-20b-code-instruct-8k$0.1/M in$0.10 per million input tokens

Per million output tokens16

PlatformModel · what it coversUSDPlatform’s own wordingSince
OpenRouterThird-partyIBM: Granite 4.2 8Bvia CoreWeave · bf16$0.15/M outCheapestcompletion: 0.00000015 / token (endpoint coreweave/bf16)
ReplicatePlatform-hostedibm-granite/granite-3.0-2b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-3.0-8b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-3.1-2b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-3.1-8b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-3.2-8b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-3.3-8b-instruct$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-4.0-h-small$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-4.1-8b$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-4.2-8b$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-8b-code-instruct-128k$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-speech-4.1-2b$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-vision-3.3-2b$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-vision-4.1-4b$0.25/M out$0.25 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-20b-code-instruct-8k$0.5/M out$0.50 per million output tokens
ReplicatePlatform-hostedibm-granite/granite-speech-3.3-8b$4.9/M out$4.90 per million output tokens

Per million cached input tokens (read)1

PlatformModel · what it coversUSDPlatform’s own wordingSince
OpenRouterThird-partyIBM: Granite 4.2 8Bvia CoreWeave · bf16$0.015/M cache readCheapestinput_cache_read: 0.000000015 / token (endpoint coreweave/bf16)

Per second of compute time1

PlatformModel · what it coversUSDPlatform’s own wordingSince
ReplicatePlatform-hostedibm-granite/granite-embedding-small-english-r2second of run time$0.00023/compute sCheapest$0.225 per thousand seconds of run time

Per run (default parameters)1

PlatformModel · what it coversUSDPlatform’s own wordingSince
ReplicatePlatform-hostedibm-granite/granite-embedding-278m-multilingualrun$0.001/runCheapest$1 per thousand runs