# Granite 4.2 8B Benchmark Scores & Performance

> Granite 4.2 8B by IBM scores 33.4/100 overall, ranking #170 out of 950 AI models.

Canonical page: https://benchlm.ai/models/granite-4-2-8b

Last updated: October 15, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | IBM |
| Source Type | Open Weight |
| Reasoning Type | Reasoning |
| Context Window | 128K |
| Official model card | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| Overall Score | 33.4/100 |
| Overall Rank | #170 of 950 |

## Family & Coverage

- Family: Granite 4.2
- Variant: 8b-dense (8B)
- Benchmarks covered: 28 of 638
- Sibling models: [Granite 4.2 30B](/models/granite-4-2-30b), [Granite 4.2 3B](/models/granite-4-2-3b)
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Agentic Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 20.6% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [τ³-bench results](/benchmarks/tau3-bench) | 58.1% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [BFCL v4](/benchmarks/bfcl-v4) | 52.4% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [GDPval-AA](/benchmarks/gdpvalaanormalized) | 0.0% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA Agentic Index](/benchmarks/aaagenticindex) | 3.7% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [GDPval-AA](/benchmarks/gdpvalaa) | 646 | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |

## Coding Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [SWE-bench Verified](/benchmarks/swe-bench-verified) | 47.7% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [SWE-bench Pro](/benchmarks/swe-bench-pro) | 19.1% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [SWE Multilingual](/benchmarks/swe-bench-multilingual) | 30.8% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 20.6% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [LiveCodeBench v6](/benchmarks/livecodebench-v6) | 73.2% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [SciCode](/benchmarks/scicode) | 36.1% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [AA-SciCode](/benchmarks/aascicode) | 31.5% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA Coding Index](/benchmarks/aacodingindex) | 22.4% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |

## Reasoning Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [AA-LCR](/benchmarks/lcr) | 45.0% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [CritPt](/benchmarks/critpt) | 0.3% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |

## Knowledge Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [GPQA](/benchmarks/gpqa) | 64.1% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [MMLU-Pro](/benchmarks/mmlu-pro) | 74.0% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [Artificial Analysis Intelligence Index](/benchmarks/artificialanalysis) | 11.1% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA-GPQA Diamond](/benchmarks/gpqa-diamond) | 63.1% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA-HLE](/benchmarks/aahle) | 9.7% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA-Omniscience Index](/benchmarks/aaomniscienceindex) | -17.2% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA-Omniscience Accuracy](/benchmarks/omniscienceaccuracy) | 11.2% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA-Omniscience Hallucination Rate](/benchmarks/omnisciencehallucinationrate) | 32.0% | [Artificial Analysis model benchmarks](https://artificialanalysis.ai/models/granite-4-2-8b) |
| [AA Openness Index](/benchmarks/aaopennessindex) | 72.2% | [Artificial Analysis: artificial-analysis-openness-index leaderboard](https://artificialanalysis.ai/evaluations/artificial-analysis-openness-index) |

## Instruction Following Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [IFBench](/benchmarks/ifbench) | 79.3% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |

## Mathematics Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [AIME 2025](/benchmarks/aime2025) | 86.7% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |
| [HMMT Feb 2025](/benchmarks/hmmtfeb2025) | 78.3% | [IBM Granite 4.2 8B model card](https://huggingface.co/ibm-granite/granite-4.2-8b) |

## Other IBM Models

- [Granite-4.0-H-1B](/models/granite-4-0-h-1b) - Score: 28.84
- [Granite-4.0-350M](/models/granite-4-0-350m) - Score: 28.48
- [Granite-4.0-H-350M](/models/granite-4-0-h-350m) - Score: 28.48
- [Granite 4.2 30B](/models/granite-4-2-30b) - Score: not computed
- [Granite 4.2 3B](/models/granite-4-2-3b) - Score: not computed
- [Granite-4.0-1B](/models/granite-4-0-1b) - Score: not computed
- [Granite Speech 5.0 470M TurboCTC](/models/granite-speech-5-0-470m-turboctc) - Score: not computed