Gemma 3 4B IT VRAM & Memory Requirements

Memory layers in this estimate

Artifact bytes, raw KV, engine extras, and total demand are separate. This is not a runtime trace.

Weight artifact

2.94 GiB (3,155,051,328 bytes)

Raw KV cache

0.27 GiB (289,406,976 bytes)

Engine extras / workspace

0.13 GiB–0.58 GiB

Total estimated demand

3.33 GiB–3.79 GiB

Published artifacts

Memory by quantization and context

Concurrent sequences = 1. KV defaults to FP16. The range includes documented runtime extras for the listed engine. These figures use the same core as the interactive calculator.

Raw KV contribution

Related pages

• KV cache calculator
• 4-bit, 8-bit, and FP16 memory
• File size is not runtime VRAM
• Methodology
• All can-i-run checks

Sources

• Gemma 3 model card (Google AI for Developers) — Input context 128K tokens for 4B, 12B, 27B and 32K for 1B; images encoded to 256 tokens each. License: Gemma Terms of Use.
• Google gemma_pytorch config.py (architecture) — get_config_for_4b: num_hidden_layers=34, num_attention_heads=8, num_key_value_heads=4, hidden_size=2560, head_dim=256, sliding_window_size=1024, attn_types = 5 LOCAL_SLIDING then 1 GLOBAL, cycled by layer index (gemma/model.py: attn_types[i % len(attn_types)]), so 5 global and 29 sliding layers. The Hugging Face config.json is gated (HTTP 401), so it was not read.
• Cross-check only: llama.cpp org GGUF header converted from Google's config (not an official Google source) — Header read with an HTTP range request on 2026-10-18: block_count 34, head_count 8, head_count_kv 4, embedding_length 2560, key_length 256, sliding_window 1024, context_length 131072. All agree with gemma_pytorch config.py.
• Gemma releases — March 10, 2025: release of Gemma 3 in 1B, 4B, 12B and 27B sizes.
• Hugging Face API (BF16) — sha 093f9f388b31de276ce2de164bdc2081324b9767; gated manual; safetensors.total=4300079472; 2 safetensors siblings.
• Hugging Face API (QAT Q4_0 GGUF) — sha 15f73f5eee9c28f53afefef5723e29680c2fc78a; gemma-3-4b-it-q4_0.gguf = 3,155,051,328 bytes.

Assumptions that affect this result

• Requested context is within the recorded native window of 131072 tokens.
• Full-resident weight estimates use total parameters, including inactive MoE experts.
• Weight bytes come from the selected weight files only, not from every file in the repository.
• Download file size is not peak GPU memory. Runtime layout, allocator padding, and KV are extra.
• Counted files: gemma-3-4b-it-q4_0.gguf.
• This checkpoint has a vision component. Text-only inference does not include vision encoder or projector bytes.
• Context budget is the number of cached tokens per sequence (prompt plus reserved generation).
• Concurrent sequences share one model replica; weights are not multiplied by sequence count.

Calculator