IBM
Granite 4.1 3B
3,4B
Parameters
2,0 Go
Q4_K_M file
≥ 3,5 Go
Recommended VRAM
131 072
Max context
Speed by graphics card
Simulate in the “Configurator” →Standard llama-bench run (prompt 512 / generation 128 tokens), model fully in VRAM. Click a card for details. Electricity cost at the default price (0,25 €/kWh).
| Card · Q4_K_M | VRAM | Generation | Prompt | VRAM used | Power | € elec. / M tok | Price |
|---|---|---|---|---|---|---|---|
| RTX 4070 Ti | 12 Go | 166 tok/s | 10 529 | 2,8 Go | 128 W | 0,05 €/M tok | — |
Not measured on these cards yet — click for an estimate:
Technical sheet
- Published
- 2026-04-06
- Architecture
- 40 layers · 8 KV heads · dimension 64
- Attention
- Standard: every layer keeps the whole context in memory.
- Context cache (f16)
- ≈ 78 MB per 1,000 tokens · 2,5 Go for 32,768 tokens
- GGUF file
- bartowski/ibm-granite_granite-4.1-3b-GGUF · ibm-granite_granite-4.1-3b-Q4_K_M.gguf Download (2,0 Go)
- Official repo
- ibm-granite/granite-4.1-3b
Public Hugging Face data (official config.json, GGUF repo). The context cache only counts full-attention layers.