whatmodel.sh / Granite 3.3 8B Instruct

Granite 3.3 8B Instruct

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Granite 3.3 8B Instruct is an open-weights language model from IBM, released April 2025. Its published size is 8.17B parameters and a 128k-token context window. Licence: apache-2.0.

Catalogue identity notes

Maker repo ibm-granite/granite-3.3-8b-instruct @51dd4bc: licence tag apache-2.0. 8,170,864,640 parameters from the Hugging Face safetensors metadata. Dense, so active = total. Released 2025-04: month the maker's Hugging Face repo was created (createdAt 2025-04-09). Ollama tags granite3.3:8b are the library's instruct/chat builds of this checkpoint (tag names on the pinned tags page); download sizes are estimated from the parameter count. The granite3.3 tags page lists only 2b (1.5GB) and 8b (4.9GB) builds, sizes consistent with 4-bit Q4_K_M; quant inferred from size.

45.6±1.2#431 of 719 LLMsfrom 13 benchmark results across 4 sources

$ how it compares

The LLMs ranked either side of Granite 3.3 8B Instruct, out of 828 with a score, then more from IBM.

#modelscorewhereRAMreleased
513Mistral Medium (2312) Mistral46±1comparison only2023-12
514GPT-3.5 Turbo (0613) OpenAI46±1comparison only2023-06
515Granite 3.3 8B Instruct IBM46±1local~6.3GB2025-04
516Tulu 2 70B Allen Institute for AI46±2local~28GB2023-11
517LFM2 8B-A1B Liquid AI46±1comparison only2025-10
more from IBM
192Granite 4.2 30B IBM63±1api $0.28/M +local~19GB2026-08
239Granite 4.2 8B IBM60±1api $0.11/M +local~6.6GB2026-08
276Granite 4.2 3B IBM57±1api $0.052/M +local~3.7GB2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #265 of 393 * · 11% of its score · 2 tests

Beats Ministral 3 14B Reasoning 2512, Ministral 3 14B, Phi-4 14B here, though they rank higher overall. Trails Flan-T5 XXL, Granite 4.0 H 1B, Granite 4.0 1B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Arena-Hard (version/judge unspecified) lab57.6% ✎█·▄··▄·····█ beats 5/5medium · avg 58%▇▇ 6.2%at its level
IFEval (aggregation unspecified) lab74.8% ✎▄▄·▄······▄█ beats 5/5easy · avg 79%▇▇ 4.9%at its level
math▁▂▃▅▆▅▄█▄▁▁▁ #257 of 398 · 45% of its score · 5 tests

Beats Molmo 2 8B, Llama 3.3 70B Instruct, Mistral Large 2 (2407) here, though they rank higher overall. Trails DeepSeek R1 Distill Qwen 1.5B, Mixtral 8x7B v0.1 (base), Tulu 2 DPO 70B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
AIME 2024 lab8.1% ✎█···▄▄▄····▄ beats 5/5very hard · avg 11%▇▇▇▇▇▇ 16%sharp · too hard for it
GSM8K lab80.9% ✎█··▄·····▄·█ beats 4/5very easy · avg 90%▇▇▇▇▇▇ 14%sharp
MATH-500 lab69.0% ✎█···▄·▄▄···▄ beats 5/5easy · avg 74%▇▇▇ 8.0%at its level
OMEGA (math generalisation)10.7% ✎█···▄·▄·▄··█ beats 2/6hard · avg 20%▇ 3.6%typical
MATH (subset unspecified)67.3% ✎██····██·███ beats 2/6easy · avg 71%▇ 3.2%at its level · noisy
other tests6 not in a task · 44% of its score
testresultvs the others in its rundifficultycountswhy
BIG-Bench Hard lab69.1% ✎█·▄·····▄▄·▄ beats 3/5easy · avg 84%▇▇▇▇▇▇ 14%sharp · at its level
MMLU lab65.5% ✎▄█········▄█ beats 3/5easy · avg 78%▇▇▇▇▇ 12%at its level
DROP lab59.4% ✎▃▃·····▃···█ beats 5/5easy · avg 68%▇▇▇▇ 10%at its level
PopQA lab26.2% ✎▄·█······▄·█ beats 3/5hard · avg 34%▇▇ 5.1%blunt · at its level
DiagnosticIQ industrial maintenance MCQ (macro accuracy, 4 options) lab59.5% ✎▁··▃▁▁▁▅█▅▁▅ beats 8/24medium · avg 52%▇ 2.4%blunt · at its level
Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate)89.4% ✎▁▂▁▂▅█▂▃▄▃▁▁ beats 37/98very easy · avg 90%▇ 0.0%blunt

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Granite 3.3 8B Instruct · needs ~4.9 GB · from huggingface.co/ibm-granite
    chat in the terminal
    llama cli -hf ibm-granite/granite-3.3-8b-instruct-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf ibm-granite/granite-3.3-8b-instruct-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Granite 3.3 8B Instruct · needs ~4.9 GB · from huggingface.co/ibm-granite

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/ibm-granite/granite-3.3-8b-instruct-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Granite 3.3 8B Instruct · needs ~5.4 GB · from ollama.com
    chat in the terminal
    ollama run granite3.3:8b

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Granite 3.3 8B Instruct · needs ~4.6 GB, Mac only · from huggingface.co/mlx-community
    chat in the terminal
    mlx_lm.chat --model mlx-community/granite-3.3-8b-instruct-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model mlx-community/granite-3.3-8b-instruct-4bit

Weights and documentation: huggingface.co/ibm-granite ↗

$ how the score was worked out · ibm-granite/granite-3.3-8b-in… 90%, allenai/Olmo-3-7B-Instruct mo… 7%, DiagnosticIQ: A Benchmark for… 2% and 1 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
ibm-granite/granite-3.3-8b-instruct model card · lab-reported9▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 90%Apache-2.0
allenai/Olmo-3-7B-Instruct model card2▇▇ 7%Apache-2.0
DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules (arXiv:2605.08614v1) · lab-reported1▇ 2%CC BY 4.0
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)1▇ 0%Apache-2.0
prior · pull toward the middle▇ 1%
weighted average= 45.6

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json