Granite 4.2 8B

IBM Granite · released Aug 31, 2026

Open weightsReasoningTool useStructured output

Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...

Compare this model Source listing

Specification

Context window: 131K
Max output: 118K
Knowledge cutoff: Not stated
Parameters: 8.8B
Licence: apache-2.0
Serving providers: 2
Moderated: No
Uptime: 100.0%

Intelligence

11.1

45th percentile

Coding

22.4

Coding Index

Agentic

1.3

Agentic Index

Output speed

118 t/s

Median across providers

Latency

247ms

Time to first token

Cost per task

$0.02

Estimated

Benchmarks

Where the score comes from

The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.

Evaluation scores

Percentage correct · higher is better

GPQA Diamond
63.1%
AA-LCR (long context)
45.0%
SciCode
31.5%
Humanity's Last Exam
9.7%
τ²-bench (Telecom)
7.6%

An evaluation missing from this list was not run for this model — it is not a zero.

View as table

Evaluation	Score
GPQA Diamond	63.1%
AA-LCR (long context)	45.0%
SciCode	31.5%
Humanity's Last Exam	9.7%
τ²-bench (Telecom)	7.6%

Against its peers

Intelligence Index · this model highlighted, nearest peers in grey

gpt-oss-120bOpenAI
11.6
Qwen3.5-9BQwen
11.2
Granite 4.2 8BIBM Granite
11.1
GLM 4.5 AirZ.ai
11.1
Qwen3 VL 32B InstructQwen
11.0
o3 Mini HighOpenAI
11.0
Trinity Large ThinkingArcee AI
10.8
Mercury 2Inception
10.8

Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.

View as table

Model	Intelligence
gpt-oss-120b	11.6
Qwen3.5-9B	11.2
Granite 4.2 8B	11.1
GLM 4.5 Air	11.1
Qwen3 VL 32B Instruct	11.0
o3 Mini High	11.0
Trinity Large Thinking	10.8
Mercury 2	10.8

Percentile among all indexed models

Intelligence45th

Coding26th

Agentic11th

Pricing

What it costs to run

List prices per million tokens, plus what one representative task works out to.

List price

Input / 1M tokens: $0.06
Output / 1M tokens: $0.25
Cached input / 1M: $0.015
Blended 3:1: $0.107

One task, estimated

$0.02

Input tokens: 50,000
Output tokens: 80,000
Profile: Reasoning

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).

Serving providers

Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.