Autor: Jakub Rusinowski · Ostatnia aktualizacja: 29 kwietnia 2026
Wydanie IBM z 29 kwietnia 2026. Warianty 3B i 8B to gęste transformery; wariant 30B wraca do hybrydowego projektu Mamba-Transformer (przeplatane warstwy SSM i uwagi) zamiast MoE z wersji 4.0. Licencja Apache 2.0, certyfikat ISO 42001, kontekst 128K (rozszerzalny do 512K). Wszystkie trzy rozmiary mają oficjalne tagi w Ollama i pokrywają zakres od budżetowego sprzętu edge po średniopółkowe GPU 24 GB.
| Licence | What it permits | Applies to |
|---|---|---|
Apache-2.0 | Commercial use permitted Commercial use permitted. No usage restrictions beyond attribution. | Granite 4.1 3B, Granite 4.1 8B, Granite 4.1 30B |
| Granite 4.1 3B | Min 3 GB VRAM · Q4_K_M · 128,000 ctx · ollama run granite4.1:3b |
| Granite 4.1 8B | Min 6 GB VRAM · Q4_K_M · 128,000 ctx · ollama run granite4.1:8b |
| Granite 4.1 30B | Min 19 GB VRAM · FP8 (Q4 est. ~12-15GB, unconfirmed) · 128,000 ctx · ollama run granite4.1:30b |
The cheapest GPU that runs IBM Granite 4.1 locally (min 3 GB VRAM) is the Intel Arc B570 (10 GB).
Install Ollama then run: ollama run granite4.1:3b
Minimum VRAM: 3 GB. For best results use Q4_K_M quantization.
IBM Granite 4.1 needs about 3 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Granite 4.1 3B (3 GB, Q4_K_M); Granite 4.1 8B (6 GB, Q4_K_M); Granite 4.1 30B (19 GB, FP8 (Q4 est. ~12-15GB, unconfirmed)). On Apple Silicon, unified memory counts toward this requirement.
Yes — IBM Granite 4.1 runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
Q4_K_M is the best balance of quality and VRAM for IBM Granite 4.1 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run granite4.1:3b. This downloads IBM Granite 4.1 and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.