Model

Granite 3.0 8B Instruct

OPEN

IBM · Granite 3.0 family · released Oct 21, 2024

Dense 8B enterprise LLM trained on 12T tokens across 12 languages/116 code languages for RAG, tool use and summarization.

ReasoningCodingVisionFunction callingTool useAgentic

415.3

Elo · rank #329

Parameters

8.1B

Active params

8.1B (dense)

Context

4K tokens

Architecture

Dense decoder-only transformer (GQA, RoPE, SwiGLU, RMSNorm)

License

Apache 2.0

Languages

12+

API price (in/out)

No hosted API

Modalities

text

Benchmark results

Bar shows position within the tracked field; marker = field best

AGIEvalReasoning40.5%#31

best: OLMo 3-Think 32B · 88.2%

ARC-ChallengeReasoning64.2%#70

best: Llama 3.1 405B · 96.9%

BIG-Bench HardReasoning51.5%#96

best: ERNIE 4.5 300B-A47B · 94.3%

GPQA DiamondReasoning33.8%#248

best: GPT-5.6 · 94.6%

GSM8KMath69.0%#101

best: Llama 3.1 405B · 96.8%

HellaSwagReasoning82.6%#62

best: Claude 3 Opus · 95.4%

HumanEvalCoding64.6%#91

best: Claude Opus 4.5 · 99.4%

IFEvalReasoning52.3%#132

best: Gemma 4 26B A4B · 98.5%

MATH-500Math30.9%#162

best: GPT-5 · 99.4%

MBPPCoding49.6%#72

best: Llama-3.3-Nemotron-Super-49B v1 (Reasoning On) · 91.3%

MGSMMath48.2%#46

best: OpenAI o4-mini · 93.7%

MMLU-ProKnowledge34.5%#160

best: Claude Fable 5 · 91.5%

MMLUKnowledge65.8%#172

best: OpenAI o3 · 92.9%

MT-BenchHuman preference8.22#39

best: Hunyuan-Large (A52B) · 9.4

OpenBookQAReasoning46.6%#32

best: Claude 1 · 90.8%

Social IQaReasoning71.2%#11

best: Apple DCLM-Baseline 7B · 82.9%

TruthfulQAKnowledge60.3%#23

best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%

WinoGrandeReasoning77.5%#47

best: PaLM 2 · 90.9%

Run it locally

VRAM @ Q4

6 GB

VRAM @ FP16

16.2 GB

Fits on (Q4)

RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB

Throughput data unavailable.

Quantizations

GGUF Q4_K_M · AWQ · GPTQ

Granite 3.0 family

Elo progression across releases

Granite 3.0 2B InstructOct 2024181.6 Granite 3.0 8B InstructOct 2024415.3

Resources

HFibm-granite/granite-3.0-8b-instruct GHibm-granite/granite-3.0-language-models DOCSibm.com/granite/docs/models/granite APIAvailable via IBM API

API price weights · each benchmark row carries its own source badge (see methodology)