Models › gemma-4-E4B-it

gemma-4-E4B-it: VRAM needs and speed by GPU

At Q4_K_M, gemma-4-E4B-it needs about 6.0 GB of VRAM. No measured runs yet. Speeds below are estimates. Submit a run.

7.52B parameters · license: apache-2.0 · 131,072-token context · GGUF published Apr 1, 2026 · GGUF files on Hugging Face

Memory needed per quantization

QuantFile sizeNeeds (estimate)Smallest card it fits
UD-IQ2_M3.5 GB4.5 GB8 GB cards
UD-IQ3_XXS3.7 GB4.7 GB8 GB cards
UD-Q2_K_XL3.8 GB4.7 GB8 GB cards
Q3_K_S3.9 GB4.9 GB8 GB cards
Q3_K_M4.1 GB5.0 GB8 GB cards
UD-Q3_K_XL4.6 GB5.6 GB8 GB cards
IQ4_XS4.7 GB5.7 GB8 GB cards
IQ4_NL4.8 GB5.8 GB8 GB cards
Q4_04.8 GB5.8 GB8 GB cards
Q4_K_S4.8 GB5.8 GB8 GB cards
Q4_K_M5.0 GB6.0 GB8 GB cards
Q4_15.1 GB6.1 GB8 GB cards
UD-Q4_K_XL5.1 GB6.1 GB8 GB cards
Q5_K_S5.4 GB6.4 GB8 GB cards
Q5_K_M5.5 GB6.5 GB8 GB cards
UD-Q5_K_XL6.7 GB7.6 GB12 GB cards
Q6_K7.1 GB8.1 GB12 GB cards
UD-Q6_K_XL7.5 GB8.4 GB12 GB cards
Q8_08.2 GB9.2 GB12 GB cards
UD-Q8_K_XL8.7 GB9.7 GB12 GB cards
BF1615.1 GB16.0 GB24 GB cards

Expected speed on popular hardware

HardwareBest quant that fitsExpected tok/s
NVIDIA RTX 3060Q8_037.6 estimate
NVIDIA RTX 4060 TiQ5_K_M44.7 estimate
NVIDIA RTX 3090Q8_080.1 estimate
NVIDIA RTX 4090Q8_088.7 estimate
NVIDIA RTX 5090Q8_0140 estimate
AMD Radeon RX 7900 XTXQ8_089.1 estimate
Apple M4 Pro (20-core GPU) (64 GB)Q8_028.3 estimate
Apple M3 Max (40-core GPU) (64 GB)Q8_041.3 estimate
AMD Ryzen AI Max+ 395 (Radeon 8060S)Q8_025.4 estimate

Check your own card

Too big for your card?

Run gemma-4-E4B-it on a rented 24-80 GB GPU by the hour.

Rent a bigger GPU on RunPod

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.