Granite 4.2 8B
IBM131K context
As of Nov 11, 2026, Granite 4.2 8B costs $0.06 per million input tokens and $0.15 per million output tokens from CoreWeave (W&B Inference). The cheapest of 3 providers is CoreWeave (W&B Inference) at $0.06 / $0.15. Price last changed Nov 7, 2026: CoreWeave (W&B Inference) input $0.10 → $0.06. Recent price changes
Providers (3)
Sorted by your monthly estimate. Prices per million tokens.
CoreWeave (W&B Inference)Cheapest
- Input / output
- $0.06 / $0.15
- Context
- 131K
- Per month
- $0.90
wandb/ibm-granite/granite-4.2-8b
- Input / output
- $0.06 / $0.25
- Context
- 131K
- Per month
- $1.10
ibm-granite/granite-4.2-8b
- Input / output
- $0.06 / $0.25
- Context
- —
- Per month
- $1.10
replicate/ibm-granite/granite-4.2-8b
Cloud and hosted providers may add regional pricing, minimums or platform fees. Supports: Reasoning, Tool calling.
Price history
CoreWeave (W&B Inference), per million tokens
Dates are when Appndo detected the change, which can be up to a day after the provider's.
Similar models
GPT-5 nano
Qwen3 30B A3B
Claude 3.7 SonnetCommon questions
How much does the Granite 4.2 8B API cost?
As of Nov 11, 2026, Granite 4.2 8B costs $0.06 per million input tokens and $0.15 per million output tokens from CoreWeave (W&B Inference). Cached input costs $0.015 per million tokens.
Where is Granite 4.2 8B cheapest?
As of Nov 11, 2026, the cheapest of 3 providers is CoreWeave (W&B Inference) at $0.06 per million input tokens and $0.15 per million output tokens. Appndo lists no official price from IBM.
What is the context length of Granite 4.2 8B?
Granite 4.2 8B takes up to 131K tokens of context at CoreWeave (W&B Inference); the limit differs between providers.
How does Granite 4.2 8B compare with similar models?
On LMArena Granite 4.2 8B scores 1319, #242 overall. Models with close scores and their cheapest price per million input / output tokens: GPT-5 nano ($0.05 / $0.40), Qwen3 30B A3B ($0.09 / $0.20), and Claude 3.7 Sonnet ($3.00 / $15.00).