Yz

granite4.1

IBM Granite Models are a family of enterprise-ready, open foundation models that support multilingual capabilities, coding, retrieval-augmented generation (RAG), tool use, and structured JSON output. Released under Apache 2.0 license.

Araç Kullanımı 3b 8b 30b
Hızlı Kurulum (Ollama kuruluysa) 
ollama run granite4.1

Ollama kurulu değil mi? ollama.com/download — Windows, macOS ve Linux için ücretsiz. İlk çalıştırmada model indirilir, sonrası tamamen çevrimdışıdır.

Varyantlar

Boyut büyüdükçe kalite artar, donanım ihtiyacı yükselir. Başlangıç için küçük varyantı deneyin.

EtiketBoyutBağlamGirdiKomut
3b 2.1GB 128K Text ollama run granite4.1:3b
8b 5.3GB 128K Text ollama run granite4.1:8b
30b 17GB 128K Text ollama run granite4.1:30b
3b-q2_K 1.4GB 128K Text ollama run granite4.1:3b-q2_K
3b-q3_K_S 1.6GB 128K Text ollama run granite4.1:3b-q3_K_S
3b-q3_K_M 1.7GB 128K Text ollama run granite4.1:3b-q3_K_M
3b-q3_K_L 1.9GB 128K Text ollama run granite4.1:3b-q3_K_L
3b-q4_0 2.0GB 128K Text ollama run granite4.1:3b-q4_0
3b-q4_1 2.2GB 128K Text ollama run granite4.1:3b-q4_1
3b-q4_K_S 2.0GB 128K Text ollama run granite4.1:3b-q4_K_S
3b-q4_K_M 2.1GB 128K Text ollama run granite4.1:3b-q4_K_M
3b-q5_0 2.4GB 128K Text ollama run granite4.1:3b-q5_0
3b-q5_1 2.6GB 128K Text ollama run granite4.1:3b-q5_1
3b-q5_K_S 2.4GB 128K Text ollama run granite4.1:3b-q5_K_S
3b-q5_K_M 2.4GB 128K Text ollama run granite4.1:3b-q5_K_M
3b-q6_K 2.8GB 128K Text ollama run granite4.1:3b-q6_K
3b-q8_0 3.6GB 128K Text ollama run granite4.1:3b-q8_0
3b-bf16 6.8GB 128K Text ollama run granite4.1:3b-bf16
8b-q2_K 3.4GB 128K Text ollama run granite4.1:8b-q2_K
8b-q3_K_S 3.9GB 128K Text ollama run granite4.1:8b-q3_K_S
8b-q3_K_M 4.3GB 128K Text ollama run granite4.1:8b-q3_K_M
8b-q3_K_L 4.7GB 128K Text ollama run granite4.1:8b-q3_K_L
8b-q4_0 5.1GB 128K Text ollama run granite4.1:8b-q4_0
8b-q4_1 5.6GB 128K Text ollama run granite4.1:8b-q4_1
8b-q4_K_S 5.1GB 128K Text ollama run granite4.1:8b-q4_K_S

23 varyant daha var — tümü için ollama.com.

Model Detayları ve Benchmarklar (kaynak: ollama.com)

Granite 4 models

Granite 4.1 language models are a family of state-of-the-art open foundation models featuring dense decoder-only architectures in three sizes — 3B, 8B, and 30B. They natively support multilingual capabilities, a wide range of coding tasks, retrieval-augmented generation (RAG), tool usage, and structured JSON output.

Our models are trained from scratch on approximately 15 trillion tokens through a five-phase strategy designed to progressively refine data quality and model capabilities. The first two phases cover pre-training proper, before transitioning into mid-training in Phases 3 and 4 with high-quality data annealing. The fifth and final phase performs long-context extension, scaling the context window up to 512K tokens through a staged process.

All models are publicly released under the Apache 2.0 license, allowing free use for both research and commercial purposes. The data curation and training processes were specifically designed for enterprise scenarios and customization, incorporating governance, risk, and compliance (GRC) evaluations alongside IBM’s standard data clearance and document quality review procedures.

We provide both base models (checkpoints after pre-training) and instruct models (checkpoints fine-tuned for dialogue, instruction following, helpfulness, and safety).

  • Please note that instruct models do not have the base qualifier in their name (e.g., ibm-granite/granite4.1:8b vs. ibm-granite/granite4.1:8b-base).

Running

Example of running the default model (i.e., typically, this is the Q4_K_M quantization):

ollama run ibm/granite4.1:8b

To run other quantizations (e.g., Q8_0):

ollama run ibm/granite4.1:8b-q8_0

Supported Languages

Supported Languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese. Users may finetune Granite 4.1 models for languages beyond these languages.

Intended Use

This model is designed to handle general instruction-following tasks and can be integrated into AI assistants across various domains, including business applications.

The model is designed to respond to general instructions and can be used to build AI assistants for multiple domains, including business applications.

Capabilities

  • Summarization
  • Text classification
  • Text extraction
  • Question-answering
  • Retrieval Augmented Generation (RAG)
  • Code related tasks
  • Function-calling tasks
  • Multilingual dialog use cases
  • Fill-In-the-Middle (FIM) code completions

Learn more