Google

Gemma 3 270M

vision

0,3B

Parameters

0,2 Go

Q4_K_M file

≥ 1,7 Go

Recommended VRAM

32 768

Max context

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_MVRAMGenerationPromptVRAM usedPower€ elec. / M tokPrice
RTX 4070 Ti12 Go728 tok/s71 2591,1 Go16 W0,00 €/M tok—

Not measured on these cards yet — click for an estimate:

Technical sheet

Published
2025-07-30
Architecture
18 layers · 1 KV heads · dimension 256
Attention
Sliding window: 3 of 18 layers keep the whole context, the others only the last 512 tokens — a much lighter cache.
Context cache (f16)
≈ 3 MB per 1,000 tokens · 0,1 Go for 32,768 tokens
GGUF file
bartowski/google_gemma-3-270m-it-GGUF · google_gemma-3-270m-it-Q4_K_M.gguf Download (0,2 Go)

Public Hugging Face data (official config.json, GGUF repo). The context cache only counts full-attention layers.