[whatmodel.sh](../../README.md) / Granite 4.0 Micro 3B
# Granite 4.0 Micro 3B
IBM · released 2025-09 · 3.4B params · apache-2.0
**score 47.1** ±0.8 · #430 of 732 · from 27 benchmark results across 4 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Granite 4.0 Micro 3B among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▄·▄········● #2 of 4* 1.2% 1
coding ●···▄······▄ #3 of 4 30% 7
chat & writing ▁▁▁▃▆▄▆●▆▇▇▃ #229 of 402* 4.5% 2
math ▁▁▃▅●▆▄█▄▂▁▁ #273 of 431 25% 6
agent tasks · #2 of 4 · 1.2% of its score
Beats [Granite 4.0 H Tiny (7B-A1B)](granite-4-0-h-tiny.md), [Granite 4.0 H Micro 3B](granite-4-0-h-micro.md) here, though they rank higher overall.
test result vs the others in its run difficulty counts why
BFCL v3 (maker) 60.0% ▁▁▁▁▁▃▅█●▅▄▁ beats 99/171 medium · avg 60% 1.2% blunt · at its level
coding · #3 of 4 · 30% of its score
Beats [Granite 4.0 H Micro 3B](granite-4-0-h-micro.md) here, though they rank higher overall.
test result vs the others in its run difficulty counts why
HumanEval (maker) 80.0% ▁▁▁▁▂▂▆▆█●▂▁ beats 176/205 easy · avg 84% 9.1% sharp
HumanEval+ (maker) 72.0% ▁·▁▁▂▃▄▄●▅▄▃ beats 90/145 easy · avg 79% 8.1% sharp · at its level
MBPP (maker) 72.0% ▁▁▁▄▅█▆▅●▁▂▁ beats 72/86 easy · avg 78% 3.7% at its level
MultiPL-E (maker) 49.2% ▃··▄·●▅▇▄█▃▄ beats 5/31 medium · avg 56% 2.8% at its level
BigCodeBench (maker) 39.2% ▅····▃·█▃▅●▃ beats 10/12 medium · avg 41% 2.8% at its level
MBPP+ (maker) 64.0% ▁··▁▁▃▄●▆▄▁▁ beats 60/120 easy · avg 68% 2.5% at its level
CRUXEval-O (maker) 41.5% ▂▂█●▅▂▁·▁·▁▁ beats 33/51 medium · avg 46% 1.5% blunt · at its level
chat & writing · #229 of 402 · 4.5% of its score
Beats [Granite 4.2 8B](granite-4-2-8b.md), [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md), [Olmo 3.1 32B Think](olmo-3-1-32b-think.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [Gemma 3 4B](gemma-3-4b.md), [Olmo 3 7B Instruct](olmo-3-7b-instruct.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
IFEval instruction-level strict (… 85.5% ▄▄····▄▄●▄·█ beats 5/8 very easy · avg 88% 2.3% typical
IFEval prompt-level strict (maker) 79.1% ▁▁··▂··▂●█▃▃ beats 8/26 easy · avg 80% 2.2% typical
math · #273 of 431 · 25% of its score
Beats [GPT-4 Turbo (2024-04-09)](gpt-4-turbo-2024-04-09.md), [Molmo 2 8B](molmo-2-8b.md), [Llama 3.3 70B Instruct](llama-3-3-70b.md) here, though they rank higher overall. Trails [DeepSeek R1 Distill Qwen 1.5B](deepseek-r1-distill-qwen-1-5b.md), [Gemma 7B (base)](gemma-7b-base.md), [CodeGemma 7B (v1.0)](codegemma-7b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
GSM8K (maker) 85.5% ▂▂▂▂▃▄▅▇█●▆▄ beats 155/206 very easy · avg 86% 8.5% sharp
Minerva MATH (maker) 62.1% ▁·▃▁▁▂▁█▄●▃▃ beats 21/27 easy · avg 68% 5.1% at its level
DeepMind Mathematics (maker) 44.6% ▅····▅·●▃▃·▅ beats 5/10 medium · avg 53% 4.7% at its level
GSM-Symbolic (maker) 79.8% ▃▃····▃·▅●█▃ beats 6/10 easy · avg 82% 4.0% typical
MGSM, 5 languages (en, es, fr, ja… 28.6% █·▄●██···█·▄ beats 3/10 medium · avg 44% 1.6% at its level
OTIS mock AIME 2024–2025 2.8% ▃●▄▃▃▃▇▇█▃▂▅ beats 23/197 very hard · avg 7.2% 0.9% sharp · too hard for it · noisy
other tests (11)
test result vs the others in its run difficulty counts why
BIG-Bench Hard (maker) 72.5% ▁▁▁▁▂▄▅▆●▃▃▁ beats 126/155 easy · avg 74% 6.6% sharp · at its level
MMLU (maker) 66.0% ▁▁▁▁▁▂▂▅●█▆▄ beats 112/214 medium · avg 73% 6.2% at its level
GPQA (subset unspecified) (maker) 30.1% ▂▁▂▂●▃▆▃▅█▃▃ beats 22/106 very hard · avg 35% 5.8% sharp
GPQA Diamond 28.3% ▁▁▁●▂▄▆▄▅▇█▅ beats 14/211 hard · avg 39% 5.5% sharp
MMLU-Pro (maker) 44.5% ▁▁·▁▁▁●▂▄▅█▁ beats 38/188 medium · avg 50% 4.4% at its level
INCLUDE (maker) 51.6% ▂····▂●▄▄▄█▄ beats 6/36 medium · avg 57% 3.7% at its level
MMMLU (maker) 55.1% ▂▁▁▂▃▂●▄▅██▃ beats 11/46 medium · avg 62% 3.4% at its level
EuroEval HellaSwag (English), tes… 70.8% ▃·▁▁▂▁▂▅●▇▄▁ beats 52/100 medium · avg 67% 1.4% sharp · at its level · noisy
EuroEval Life in the UK (English … 77.3% ▁▁▁▁▁▁▁▂●█▄▃ beats 57/146 easy · avg 82% 1.0% at its level · noisy
Vectara hallucination leaderboard… 93.5% ▁▁▂▃▄●█▆▃▂▂▁ beats 47/140 very easy · avg 94% 0.5% blunt · maxed out · noisy
Chess Puzzles (Epoch) 0.0% ●▁▁▁▁▂▃▃▃▂▁▁ beats 45/137 very hard · avg 5.8% 0.2% sharp · too hard for it · noisy
source results share of the score
ibm-granite/granite-4.0-h-small model c… 21 90% ▇▇▇▇▇▇▇▇▇▇▇
Epoch AI 3 7% ▇
EuroEval leaderboards 2 2% ▇
Vectara hallucination leaderboard 1 1% ▇
prior: pull toward the middle 0% ▇
weighted average = 47.1
Licences: ibm-granite/granite-4.0-h-small model card ([Apache-2.0](../../licenses/ibm-granite--granite-4.0-h-small.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Granite 4.0 Micro 3B (Q4_K_M, needs ~2.1 GB, from [huggingface.co/ibm-granite](https://huggingface.co/ibm-granite/granite-4.0-micro-GGUF/tree/ec48475f0c811d812fbfb61975717a9c36eeb652))
```sh
llama cli -hf ibm-granite/granite-4.0-micro-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf ibm-granite/granite-4.0-micro-GGUF:Q4_K_M
```
Other versions: Q8_0 (~3.6 GB) `llama cli -hf ibm-granite/granite-4.0-micro-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Granite 4.0 Micro 3B (Q4_K_M, needs ~2.1 GB, from [huggingface.co/ibm-granite](https://huggingface.co/ibm-granite/granite-4.0-micro-GGUF/tree/ec48475f0c811d812fbfb61975717a9c36eeb652))
```sh
lms get https://huggingface.co/ibm-granite/granite-4.0-micro-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~3.6 GB) `lms get https://huggingface.co/ibm-granite/granite-4.0-micro-GGUF@Q8_0 -y` · 4bit (~1.9 GB, Mac only) `lms get https://huggingface.co/ethicalabs/granite-4.0-micro-MLX -y` · 8bit (~3.6 GB, Mac only) `lms get https://huggingface.co/mlx-community/granite-4.0-micro-8bit -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Granite 4.0 Micro 3B (Q4_K_M, needs ~2.54 GB, from [ollama.com](https://ollama.com/library/granite4/tags))
```sh
ollama run granite4:micro
```
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Granite 4.0 Micro 3B (4bit, needs ~1.9 GB, Mac only, from [huggingface.co/ethicalabs](https://huggingface.co/ethicalabs/granite-4.0-micro-MLX/tree/8681a64eae6f7091f34a68b93c497283c16a235a))
```sh
mlx_lm.chat --model ethicalabs/granite-4.0-micro-MLX
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model ethicalabs/granite-4.0-micro-MLX
```
Other versions: 8bit (~3.6 GB, Mac only) `mlx_lm.chat --model mlx-community/granite-4.0-micro-8bit`
Model repository and documentation: [huggingface.co/ibm-granite](https://huggingface.co/ibm-granite/granite-4.0-micro)
## Published speed measurements
Reference results for the hardware, engine and settings below. These are not verified speeds of the downloads above or predictions for your machine.
hardware precision speed engine and settings
Intel CoffeeLake-S GT2 [UHD… q4_k_m 6.18 tok/s ollama (version not recorded) · l…
Each speed links to its source, which has the full settings.
[interactive version](https://whatmodel.sh/model/granite-4-0-micro) · [JSON](https://whatmodel.sh/api?format=json&all=1)