Granite 4.2 8B

Benchmarks, price, context and versions

Also searched as

Granite-4.2-8B

Direct answer: Granite 4.2 8B is a model from IBM. 65.7% on GPQA Diamond (OpenRouter), per OpenRouter Benchmarks. On OpenRouter, US$ 0.06 per 1 million input tokens and US$ 0.15 per 1 million output tokens.

✓ Good for

high-volume use (input price in the cheapest 25% on OpenRouter)
running on your own infrastructure (open weights)

✕ Not the best for

technical reasoning as the model appears through the API (#113 of 135 on GPQA Diamond (OpenRouter))
agents that follow rules and use tools (#97 of 131 on τ-bench Verified (airline))

This block follows a rule of this section, using source data: if the model is in the top 25% of the metric, the item goes under "Good for"; if it is in the bottom half, it goes under "Not the best for". The middle band appears in neither. Price and context use OpenRouter; scores use Epoch AI or OpenRouter Benchmarks, in the latest version we checked.

Benchmarks

GPQA Diamond (OpenRouter)

?what it is

65.7%

highest score on this metric 96.5%

113º / 135 · OpenRouter Benchmarks, latest version we checked

τ-bench Verified (airline)

64%

highest score on this metric 82%

97º / 131 · OpenRouter Benchmarks, latest version we checked

Key specs

Lab: IBM
Origin: the US
Family: Granite provisional data
Weights?: open
License: no data
Input US$/1M?: US$ 0.06
Output US$/1M: US$ 0.15
Context?: 131,072 tokens
Parameters: no data
Input: text
OpenRouter usage?: no data
Downloads (30 days)?: no data
Use cases provisional data: Text Agents
Dates in sources: 2026, listed on OpenRouter

Version history

2026 granite-switch-4.1-8b Hugging Face Hub, Hugging Face repo created
2026 granite-switch-4.1-30b Hugging Face Hub, Hugging Face repo created
2026 granite-speech-5.0-2b Hugging Face Hub, Hugging Face repo created
2026 ensemble-forecasting-with-ibm-granite-time-series Hugging Face Hub, Hugging Face repo created
2026 granite-timeseries-ensemble-r1 Hugging Face Hub, Hugging Face repo created
2026 IBM: Granite 4.2 8B OpenRouter, listed on OpenRouter
2026 granite-speech-5.0-470m-turboctc-nc Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-8b-fp8 Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-3b-fp8 Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-30b-fp8 Hugging Face Hub, Hugging Face repo created
2026 granite-timeseries-patchtst-fm-r2 Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-30b Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-8b Hugging Face Hub, Hugging Face repo created
2026 granite-4.2-3b Hugging Face Hub, Hugging Face repo created

Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch.

FAQ

What is Granite 4.2 8B?

Granite 4.2 8B is a model from IBM, with open weights, used for text, agents.

How much does Granite 4.2 8B cost?

On OpenRouter, US$ 0.06 per 1 million input tokens and US$ 0.15 per 1 million output tokens.

What is the context window of Granite 4.2 8B?

131,072 tokens, per OpenRouter.

Is Granite 4.2 8B open source?

Yes, it has open weights. License: no data.

What is Granite 4.2 8B's GPQA Diamond (OpenRouter) score?

65.7% on GPQA Diamond (OpenRouter), per OpenRouter Benchmarks.

Is Granite 4.2 8B good for coding?

No data: there is no SWE-bench Verified measurement for it in open sources.

Who made Granite 4.2 8B?

It was made by IBM, a lab from the US.

provisional data Questions under review: Semrush had no API balance during collection. Answers use source data only.

Sources

OpenRouter, dados públicos da API; exports de benchmarks e rankings em CC BY 4.0, latest version we checked.
OpenRouter Benchmarks, CC BY 4.0, latest version we checked.

Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.