Inference
EN
IBM

Granite 4.2 8B

Granite 4.2 8B is a reasoning model from IBM in the Granite 4.2 family. 17 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-08
RankedEnough results for an index.
IndexAll categories, out of 100.37.6±5.4Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.101/sMedian 67/s
First tokenSeconds to the first output token.0.3 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.06Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.15Median $2.55
ContextMaximum tokens in one request.131KMedian 262K
EloLMArena rating and rank.1319±11 #242Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
ibm-granite/granite-4.2-8b
Max outputMaximum output tokens per request.
118K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.015
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
2 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
CoreWeaveHost$0.06$0.15131K
OpenRouterRouter$0.06$0.15131K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

8.79B parameters
16-bit
19.5 GB
8-bit
10.8 GB
4-bit
6.6 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 8 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted36.8185 / 218
CodingCode writing and repair.4 counted36.5188 / 209
ReasoningLogic problems and puzzles.2 counted37.8177 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.4 counted33.3232 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted63.2—
MathMath problems.2 counted41.9120 / 168

Results

17 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Instruction Following BenchmarkInstructionAbove fitFar above this model's index.79.3%16 / 57
American Invitational Mathematics Examination 2025Math86.7%24 / 53
τ³-Bench Tool-Agent-User EvaluationAgentic58.1%16 / 32
Berkeley Function Calling Leaderboard v4Agentic52.4%16 / 26
Critical Physics TasksReasoning0.3%137 / 214
LiveCodeBench v6Coding73.2%38 / 59
Artificial Analysis Humanity's Last ExamKnowledge9.7%148 / 222
Massive Multitask Language Understanding ProfessionalKnowledge74.0%53 / 75
Graduate-Level Google-Proof Q&AKnowledge64.1%106 / 141
Artificial Analysis GPQA DiamondKnowledge63.1%150 / 199
Artificial Analysis Long Context ReasoningReasoning45.0%164 / 217
GDPval-AA normalizedAgentic0.0%101 / 133
Scientific Code BenchmarkCoding36.1%27 / 35
Harvard-MIT Mathematics Tournament February 2025Math78.3%13 / 15
Software Engineering Benchmark VerifiedCoding47.7%101 / 114
Artificial Analysis SciCodeCoding31.5%104 / 114
Terminal-Bench 2.1 · BenchLMAgentic20.6%83 / 86
Artificial Analysis Omniscience AccuracyKnowledge11.8%159 / 164
SWE-bench ProCoding19.1%82 / 83
Artificial Analysis Openness IndexKnowledge72.2%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reusemodels.devMIT

History

5 releases
Index38.3 → 37.6
23 Oct 202612 Nov 2026
Output / 1M$0.25 → $0.15
23 Oct 202612 Nov 2026

Questions

How much does Granite 4.2 8B cost?
$0.06 per 1M input tokens, $0.15 per 1M output tokens.
How good is Granite 4.2 8B?
Its index is 37.6 out of 100. That is place 225 of 283 ranked models.
What is Granite 4.2 8B best at?
Math: 41.9 out of 100, place 120 of 168.
How fast is Granite 4.2 8B?
101 output tokens per second, 0.3 s to the first token.
What is the context window of Granite 4.2 8B?
131K tokens.
Is Granite 4.2 8B open source?
Yes. Its weights are published.
When was Granite 4.2 8B released?
August 2026.

Badge

Granite 4.2 8B on Inference360: place 225 of 283, index 37.6
HTML
<a href="https://inference360.ai/models/granite-4-2-8b"><img src="https://inference360.ai/badges/granite-4-2-8b.svg" alt="Granite 4.2 8B on Inference360: place 225 of 283, index 37.6" height="20"></a>
Markdown
[![Granite 4.2 8B on Inference360: place 225 of 283, index 37.6](https://inference360.ai/badges/granite-4-2-8b.svg)](https://inference360.ai/models/granite-4-2-8b)

Compare with

Claude 3.7 Sonnet37.7Mistral Small 437.9DeepSeek V3 032437.1MiniCPM5-2B37.1Granite 4.2 30B41.6

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from IBM

All models
Granite 4.2 30B41.6Granite 4.2 3B33.2Granite-4.0-H-1B16.3Granite-4.0-H-350M12.7Granite-4.0-350M11.9