Skip to content
Models / Gemma 4 12B

Gemma 4 12B

consumer
HuggingFace → HuggingBay → Google → Cost & VRAM calculator → Find for my hardware

Smartest model in the 8-16GB class. Competes with models 2x larger. Gemma 4 generation with MTP support.

AI-generated content marks

The provider reports that this model does not add embedded watermarks or provenance metadata to generated output.

Provider transparency docs →

coding reasoning chat general
Parameters
12.0B
Context
128k
License
other
Developer
Google
Origin
🇺🇸 USA
Released
Jun 2025

Scores

Coding
70
Reasoning
74
General
72

Related models

Gemma 4 26B A4B
25.2B 3.8B active enthusiast 15GB min RAM $0.02/M in Apr 30, 2026
Gemma 4 31B
30.7B enthusiast 19GB min RAM $0.10/M in Apr 2, 2026
Gemma 3 270M
270M edge 1GB min RAM Aug 14, 2025
Gemma 3n E2B
2.0B edge 6GB min RAM Jun 26, 2025
Gemma 3n E4B
4.0B edge 8GB min RAM Jun 26, 2025
Gemma 3 12B
12.2B consumer 8GB min RAM $0.08/M in Feb 27, 2025
Gemma 3 1B
1.0B edge 1GB min RAM Feb 27, 2025
Gemma 3 27B
27.2B consumer 17GB min RAM $0.08/M in Feb 27, 2025
Gemma 3 4B
4.3B edge 3GB min RAM $0.04/M in Feb 27, 2025

Save your hardware and every model page answers the real question: will it run on your machine, and how fast?

Join free - save your rig →

Run it locally

Per-quant memory needs and a static "can you run it?" reference - no rig entry required

Q4_K_M
7.2GB 7.2GB min 8.9GB rec
Balanced - the usual local sweet spot
Q8_0
13.0GB 13.0GB min 14.7GB rec
Near-lossless

The reference hardware

Schematic of the 4x H100 80GB (320GB) reference rig - 320GB VRAM, 13400 GB/s aggregate bandwidth
Schematic of the NVIDIA DGX Station 748GB reference rig - 748GB unified memory, 8000 GB/s aggregate bandwidth
Schematic of the 8x RTX 3090 rack (192GB) reference rig - 192GB VRAM, 7489 GB/s aggregate bandwidth
Schematic of the 4x RTX 5090 (128GB) reference rig - 128GB VRAM, 7168 GB/s aggregate bandwidth
Schematic of the AMD Instinct MI300X (192GB) reference rig - 192GB VRAM, 5324 GB/s aggregate bandwidth
Schematic of the 4x RTX 4090 (96GB) reference rig - 96GB VRAM, 4032 GB/s aggregate bandwidth
Schematic of the 2x RTX 5090 (64GB) reference rig - 64GB VRAM, 3584 GB/s aggregate bandwidth
Schematic of the 2x RTX 3090 (48GB) reference rig - 48GB VRAM, 1872 GB/s aggregate bandwidth
Schematic of the Single RTX 5090 (32GB) reference rig - 32GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the RTX PRO 6000 Blackwell (96GB) reference rig - 96GB VRAM, 1792 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 192GB reference rig - 192GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Mac Studio M4 Ultra 512GB reference rig - 512GB unified memory, 1092 GB/s aggregate bandwidth
Schematic of the Single RTX 4090 (24GB) reference rig - 24GB VRAM, 1008 GB/s aggregate bandwidth
Schematic of the MacBook Pro M5 Max 128GB reference rig - 128GB unified memory, 614 GB/s aggregate bandwidth
Schematic of the Single GTX 1080 Ti (11GB) reference rig - 11GB VRAM, 484 GB/s aggregate bandwidth
Schematic of the Dual EPYC 9004 + 768GB DDR5-4800 reference rig - 768GB unified memory, 460 GB/s aggregate bandwidth
Schematic of the DGX Spark 128GB unified reference rig - 128GB unified memory, 273 GB/s aggregate bandwidth
Schematic of the Ryzen AI Max+ 395 128GB reference rig - 128GB unified memory, 256 GB/s aggregate bandwidth
Schematic of the Jetson AGX Orin 64GB reference rig - 64GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-3200 + 2x RTX 3090 reference rig - 560GB unified memory, 204 GB/s aggregate bandwidth
Schematic of the Epyc + 512GB DDR4-2400 + 2x RTX 3090 reference rig - 560GB unified memory, 153 GB/s aggregate bandwidth
Schematic of the NVIDIA Jetson Orin NX 16GB reference rig - 16GB unified memory, 102 GB/s aggregate bandwidth

22 reference configs, drawn in-house. Scroll for more.

Can you run it? - reference rigs

Rig Q4_K_M Q8_0
4x H100 80GB (320GB) fast 814.9t/s fast 496.5t/s
NVIDIA DGX Station 748GB fast 486.5t/s fast 296.4t/s
8x RTX 3090 rack (192GB) fast 455.5t/s fast 277.5t/s
4x RTX 5090 (128GB) fast 435.9t/s fast 265.6t/s
AMD Instinct MI300X (192GB) fast 323.8t/s fast 197.3t/s
4x RTX 4090 (96GB) fast 245.2t/s fast 149.4t/s
2x RTX 5090 (64GB) fast 218.0t/s fast 132.8t/s
2x RTX 3090 (48GB) fast 113.9t/s fast 69.4t/s
Single RTX 5090 (32GB) fast 109.0t/s fast 66.4t/s
RTX PRO 6000 Blackwell (96GB) fast 109.0t/s fast 66.4t/s
Mac Studio M4 Ultra 192GB fast 72.5t/s fast 44.1t/s
Mac Studio M4 Ultra 512GB fast 72.5t/s fast 44.1t/s
Single RTX 4090 (24GB) fast 61.3t/s fast 37.4t/s
MacBook Pro M5 Max 128GB fast 40.7t/s fast 24.8t/s
Single GTX 1080 Ti (11GB) fast 29.4t/s offload
Dual EPYC 9004 + 768GB DDR5-4800 fast 28.0t/s ok 17.1t/s
DGX Spark 128GB unified ok 16.6t/s ok 10.1t/s
Ryzen AI Max+ 395 128GB ok 15.6t/s ok 9.5t/s
Jetson AGX Orin 64GB ok 12.5t/s slow 7.6t/s
Epyc + 512GB DDR4-3200 + 2x RTX 3090 ok 12.5t/s slow 7.6t/s
Epyc + 512GB DDR4-2400 + 2x RTX 3090 ok 9.3t/s slow 5.7t/s
NVIDIA Jetson Orin NX 16GB slow 6.2t/s slow 3.8t/s

Fit tiers use the same will-it-run logic as the rig finder. For comfortable fits, the badge reflects decode speed: fast >=20 t/s, ok 8-20 t/s, slow <8 t/s. t/s is a bandwidth estimate, not a measured benchmark.

Download options

Q4_K_M official -4% vs fp16
7.2GB dl 7.2GB min 8.9GB rec
REC RAM vs largest quant
Q8_0 official -1% vs fp16
13.0GB dl 13.0GB min 14.7GB rec
REC RAM vs largest quant

Or run it in the cloud

No per-token API provider pricing tracked for Gemma 4 12B yet. For flagship list prices, see the calculator.

See who runs Google in production →

PRICE HISTORY

Inference cost over time

Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.

Loading price history...

Tokenstead community

Need a second opinion on this model?

Ask the people running it on real hardware, then post your tok/s and fit notes so the next reader has a data point.

Join the community Share your rig

Something wrong or stale? Send a correction.