Gemma 4 12B

Needs ~8.8GB of memory at Q4_0 and runs on 67 of the 99 phones we track — 0 of them smoothly.

PARAMS
12B
FAMILY
gemma
RELEASED
2026-03
TAGS
chat · vision · reasoning

Downloads & memory by quant

The download is only the weights — add KV cache and runtime to get what your phone actually needs. ★ = recommended.

QuantDownloadMemory neededFits on
Q4_07 GB~8.8 GB67 of 99 phones

Gemma 4 12B on 99 phones

At Q4_0, on each phone's largest RAM option. Tap a phone for the full report.

PhoneChip · RAMNeedsSpeedVerdict
AOPPO Find X8 ProDimensity 9400 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
Avivo X200 ProDimensity 9400 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AGalaxy S26 UltraSnapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AXiaomi 17Snapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AXiaomi 17 ProSnapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AOnePlus 15Snapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
Avivo X300 ProDimensity 9500 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AOPPO Find X9 ProDimensity 9500 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
ARedMagic 11S ProSnapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AHonor Magic8 ProSnapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AXperia 1 VIIISnapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AGalaxy Z Fold8Snapdragon 8 Elite Gen 5 · 16 GB8.8 / 12 GB~5.5 tokens/s! Runs, barely
AGalaxy S25 UltraSnapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AOnePlus 13Snapdragon 8 Elite · 24 GB8.8 / 20 GB~4.9 tokens/s! Runs, barely
AOnePlus 12Snapdragon 8 Gen 3 · 24 GB8.8 / 20 GB~4.9 tokens/s! Runs, barely
AXiaomi 15 ProSnapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AXiaomi 14Snapdragon 8 Gen 3 · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
Avivo X100 ProDimensity 9300 · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AROG Phone 9 ProSnapdragon 8 Elite · 24 GB8.8 / 20 GB~4.9 tokens/s! Runs, barely
ANothing Phone (3)Snapdragon 8s Gen 4 · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AHONOR Magic7 ProSnapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AGalaxy Z Fold7Snapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AMotorola Razr Ultra (2025)Snapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AOnePlus 13RSnapdragon 8 Gen 3 · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AOnePlus 15RSnapdragon 8 Gen 5 · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AMotorola Razr Ultra (2026)Snapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
ARedMagic 11 AirSnapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
AXperia 1 VIISnapdragon 8 Elite · 16 GB8.8 / 12 GB~4.9 tokens/s! Runs, barely
APixel 9 ProGoogle Tensor G4 · 16 GB8.8 / 12 GB~4.4 tokens/s! Runs, barely
APixel 10 ProGoogle Tensor G5 · 16 GB8.8 / 12 GB~4.4 tokens/s! Runs, barely
APixel 9 Pro XLGoogle Tensor G4 · 16 GB8.8 / 12 GB~4.4 tokens/s! Runs, barely
APixel 10 Pro XLGoogle Tensor G5 · 16 GB8.8 / 12 GB~4.4 tokens/s! Runs, barely
AMotorola Edge 50 UltraSnapdragon 8s Gen 3 · 16 GB8.8 / 12 GB~4.3 tokens/s! Runs, barely
BGalaxy S26Snapdragon 8 Elite Gen 5 · 12 GB8.8 / 9 GB~5.5 tokens/s! Runs, barely
BGalaxy S26+Snapdragon 8 Elite Gen 5 · 12 GB8.8 / 9 GB~5.5 tokens/s! Runs, barely
BXiaomi 15T ProDimensity 9400+ · 12 GB8.8 / 9 GB~5.5 tokens/s! Runs, barely
BGalaxy Z Flip8Snapdragon 8 Elite Gen 5 · 12 GB8.8 / 9 GB~5.5 tokens/s! Runs, barely
BGalaxy S25Snapdragon 8 Elite · 12 GB8.8 / 9 GB~4.9 tokens/s! Runs, barely
BGalaxy S24 UltraSnapdragon 8 Gen 3 · 12 GB8.8 / 9 GB~4.9 tokens/s! Runs, barely
BGalaxy Z Fold6Snapdragon 8 Gen 3 · 12 GB8.8 / 9 GB~4.9 tokens/s! Runs, barely
BGalaxy Z Flip6Snapdragon 8 Gen 3 · 12 GB8.8 / 9 GB~4.9 tokens/s! Runs, barely
BGalaxy Z Flip7Exynos 2500 · 12 GB8.8 / 9 GB~4.9 tokens/s! Runs, barely
CGalaxy S24Samsung Exynos 2400 · 12 GB8.8 / 9 GB~4.3 tokens/s! Runs, barely
CGalaxy S23 UltraSnapdragon 8 Gen 2 · 12 GB8.8 / 9 GB~4.3 tokens/s! Runs, barely
CPixel 9Google Tensor G4 · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CPixel 8 ProGoogle Tensor G3 · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CPOCO X6 ProDimensity 8300 Ultra · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CPixel 10Google Tensor G5 · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CMotorola Razr+ (2026)Snapdragon 8s Gen 3 · 12 GB8.8 / 9 GB~4.3 tokens/s! Runs, barely
CXiaomi 15TDimensity 8400 Ultra · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CPOCO X7 ProDimensity 8400 Ultra · 12 GB8.8 / 9 GB~4.4 tokens/s! Runs, barely
CAQUOS R10Snapdragon 7+ Gen 3 · 12 GB8.8 / 9 GB~4.3 tokens/s! Runs, barely
CGalaxy A57 5GExynos 1680 · 12 GB8.8 / 9 GB~2.2 tokens/s! Runs, barely
CNothing Phone (4a) ProSnapdragon 7 Gen 4 · 12 GB8.8 / 9 GB~2.2 tokens/s! Runs, barely
CGalaxy A55Exynos 1480 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 13 ProSnapdragon 7s Gen 2 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CNothing Phone (2a)Dimensity 7200 Pro · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CNothing Phone (3a)Snapdragon 7s Gen 3 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CGalaxy A36 5GSnapdragon 6 Gen 3 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CGalaxy A56 5GExynos 1580 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CGalaxy A37 5GExynos 1480 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 14 Pro 5GDimensity 7300 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 15 Pro 5GDimensity 7400 Ultra · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 15 5GSnapdragon 6 Gen 3 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 14 Pro+ 5GSnapdragon 7s Gen 3 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CRedmi Note 15 Pro+ 5GSnapdragon 7s Gen 4 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
CHonor 400Snapdragon 7 Gen 3 · 12 GB8.8 / 9 GB~1.6 tokens/s! Runs, barely
FiPhone AirApple A19 Pro · 12 GB8.8 / 7.8 GB Won't fit
FiPhone 17 ProApple A19 Pro · 12 GB8.8 / 7.8 GB Won't fit
FiPhone 17 Pro MaxApple A19 Pro · 12 GB8.8 / 7.8 GB Won't fit
FPixel 8aGoogle Tensor G3 · 8 GB8.8 / 6 GB Won't fit
FPixel 9aGoogle Tensor G4 · 8 GB8.8 / 6 GB Won't fit
FGalaxy S25 FESamsung Exynos 2400 · 8 GB8.8 / 6 GB Won't fit
FGalaxy A16 5GExynos 1330 · 8 GB8.8 / 6 GB Won't fit
FGalaxy A17 5GExynos 1330 · 8 GB8.8 / 6 GB Won't fit
FGalaxy A26 5GExynos 1380 · 8 GB8.8 / 6 GB Won't fit
FMoto G (2025)Dimensity 6300 · 8 GB8.8 / 6 GB Won't fit
FMoto G Power (2025)Dimensity 6300 · 8 GB8.8 / 6 GB Won't fit
FMoto G Stylus (2025)Snapdragon 6 Gen 3 · 8 GB8.8 / 6 GB Won't fit
FMoto G Power (2026)Dimensity 6300 · 8 GB8.8 / 6 GB Won't fit
FMoto G Stylus (2026)Snapdragon 6 Gen 3 · 8 GB8.8 / 6 GB Won't fit
FMotorola Razr (2025)Dimensity 7400X · 8 GB8.8 / 6 GB Won't fit
FPixel 10aGoogle Tensor G4 · 8 GB8.8 / 6 GB Won't fit
FMotorola Razr (2026)Dimensity 7450X · 8 GB8.8 / 6 GB Won't fit
FMotorola Edge (2026)Dimensity 7450 · 8 GB8.8 / 6 GB Won't fit
FCMF Phone 2 ProDimensity 7300 · 8 GB8.8 / 6 GB Won't fit
FAQUOS sense10Snapdragon 7s Gen 3 · 8 GB8.8 / 6 GB Won't fit
FAQUOS sense9Snapdragon 7s Gen 2 · 8 GB8.8 / 6 GB Won't fit
FXperia 10 VIISnapdragon 6 Gen 3 · 8 GB8.8 / 6 GB Won't fit
FiPhone 16 Pro MaxApple A18 Pro · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 16 ProApple A18 Pro · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 16Apple A18 · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 15 ProApple A17 Pro · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 17Apple A19 · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 16eApple A18 · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 17eApple A19 · 8 GB8.8 / 5.2 GB Won't fit
FiPhone 15Apple A16 Bionic · 6 GB8.8 / 3.9 GB Won't fit
FiPhone 14Apple A15 Bionic · 6 GB8.8 / 3.9 GB Won't fit
FMoto G (2026)Dimensity 6300 · 4 GB8.8 / 2 GB Won't fit

~ = bandwidth-based estimate · ✓ = measured on real hardware

FAQ

How much RAM do I need to run Gemma 4 12B on a phone?

At Q4_0, Gemma 4 12B needs ~8.8GB of usable memory (weights + KV cache + runtime). In practice that means a 12GB+ Android phone or a 16GB+ iPhone — iOS lets a single app use less of its RAM than Android does.

How fast does Gemma 4 12B run on a flagship phone?

On the OPPO Find X8 Pro (Dimensity 9400) it runs at ~5.5 tokens/s at Q4_0, estimated from memory bandwidth. Anything above ~8 tokens/s feels smooth for chat.

Can an iPhone run Gemma 4 12B?

Not really: the best iPhone we track (iPhone Air) has ~7.8GB usable, but Gemma 4 12B needs 8.8GB at Q4_0.

What is the best quantization of Gemma 4 12B for mobile?

Q4_0 (7GB download) is the size/quality sweet spot. Total memory needed is ~8.8GB once the KV cache and runtime are counted — the download size alone understates it.

How many phones can run Gemma 4 12B?

67 of the 99 phones we track fit it at Q4_0; 0 run it smoothly. The rest fall short on usable RAM.

What is Gemma 4 12B good for on a phone?

It's tagged for chat, vision, reasoning. At 12B parameters, it prioritizes answer quality over speed — expect slower decoding.

Similar models

Gemma 3 1BGemma 3 4BGemma 3 12BGemma 4 E2BGemma 4 E4B