Granite 3.3 8B Instruct on Replicate API

Granite 3 · IBM Research

ServerlessOpen Source

Last refreshed 2026-09-22. Next refresh: weekly.

Why use Granite 3.3 8B Instruct on Replicate API?

Replicate API offers Granite 3.3 8B Instruct with pay-as-you-go pricing at $0.03/1M input tokens. Replicate is a cloud-based platform that enables users to run machine learning models easily and efficiently.

Compare Granite 3.3 8B Instruct across 2 providers to find the best fit for your use case
Input / 1M
$0.030
Output / 1M
$0.25
Cache
Not sourced
Batch
Not sourced

Setup recipe

Python + curl
Install
pip install replicate
Auth
export REPLICATE_API_TOKEN=...
Call
import replicate
output = replicate.run(
    "ibm-granite/granite-3.3-8b-instruct",
    input={"prompt": "Hello"}
Model ID
ibm-granite/granite-3.3-8b-instruct

Request example

import replicate

# reads REPLICATE_API_TOKEN from env
# ibm-granite/granite-3.3-8b-instruct format: "owner/model-name" (latest version) or "owner/model-name:version-hash"
output = replicate.run(
    "ibm-granite/granite-3.3-8b-instruct",
    input={"prompt": "Hello"}
)
# Output is a list or generator depending on the model
print("".join(output))

Gotchas

  • Use provider model ID "ibm-granite/granite-3.3-8b-instruct", not the LLMReference slug "granite-3.3-8b-instruct".
  • Replicate uses "owner/model-name" format (e.g. "meta/meta-llama-3-8b-instruct") for the latest version, or "owner/model-name:version-sha" to pin to a specific version. The REST endpoint splits owner and model-name into the path: /v1/models/{owner}/{model-name}/predictions.
  • The examples expect REPLICATE_API_TOKEN; rename it only if your application config maps the new variable.

Compare Granite 3.3 8B Instruct Across Providers

ProviderInput (per 1M)Output (per 1M)
NVIDIA NIM——
Replicate API$0.03$0.25

Pricing

TypePrice (per 1M)
Input tokens$0.03
Output tokens$0.25

Capabilities

JSON / Tool use

About Granite 3.3 8B Instruct

IBM Granite 3.3 8B with improved reasoning capabilities. Part of IBM's enterprise-focused Granite model family optimized for instruction following.

Get Started

Model Specs

Released2025-03-01
Parameters8B
Context128k
ArchitectureDecoder Only
Knowledge cutoff2024-04