Skip to content

Gemma-4 E4B-it

Q4_K_M·4B params·128K ctx·GGUF
visiontool-callingtool-callingvisionllamacpp
checkpoint: unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
commit: ce152932ac27
weights 5.56 GiB · on-disk 5.00 GiB

All runs (35)

legacystack comparable
GeForce RTX 3090 · 24 GiB450 W maxdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-450wchat1
148.2
135.9608.259ms6.736100736ms0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 350 Wdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-350wchat1
147.3
137.3538.769ms6.836100728ms0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiB450 W maxdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-450wcodegen1
146.2
136.6728.298ms6.87110007.32s0.010 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiB450 W maxdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-450wrag1
145.3
118.94744.5180ms6.98532001.68s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 350 Wdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-350wcodegen1
145.1
137.3742.096ms6.97110007.29s0.010 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiB450 W maxdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-450wagent4
144.8
56.1139.45.62s6.96185009.23s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 350 Wdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-350wrag1
144.6
122.55377.9196ms6.98532001.63s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiB450 W maxdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-450wagent1
144.4
133.44191.2148ms6.96185003.75s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 350 Wdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-350wagent1
143.9
129.84493.2138ms6.96185003.85s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 350 Wdrv 590
llama.cpp cuda-4f13cb7 (cuda)baseline-pl-350wagent4
143.8
55.2136.95.68s7.06185009.39s0.000 GiB
legacystack comparable
GeForce RTX 5070 · 12 GiBcap 250 Wdrv 595
llama.cpp b9174 (vulkan)baselinechat1
129.8
124.3612.559ms7.736100804ms0.000 GiB
legacystack comparable
GeForce RTX 5070 · 12 GiBcap 250 Wdrv 595
llama.cpp b9174 (vulkan)baselinecodegen1
128.1
126.81069.167ms7.87110007.89s0.000 GiB
legacystack comparable
GeForce RTX 5070 · 12 GiBcap 250 Wdrv 595
llama.cpp b9174 (vulkan)baselinerag1
127.0
120.56792.3118ms7.98532001.66s0.000 GiB
legacystack comparable
GeForce RTX 5070 · 12 GiBcap 250 Wdrv 595
llama.cpp b9174 (vulkan)baselineagent1
124.8
123.210548.259ms8.06185004.06s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 200 Wdrv 590
llama.cpp 59778f0 (cuda)baselinechat1
123.6
118.4545.366ms8.136100845ms0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 200 Wdrv 590
llama.cpp 59778f0 (cuda)baselinerag1
119.3
101.93051.7307ms8.48532001.96s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 200 Wdrv 590
llama.cpp 59778f0 (cuda)baselinecodegen1
118.7
117.2703.3101ms8.47110008.53s0.010 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 200 Wdrv 590
llama.cpp 59778f0 (cuda)baselineagent1
116.2
111.73614.7171ms8.66185004.48s0.000 GiB
legacystack comparable
GeForce RTX 5070 · 12 GiBcap 250 Wdrv 595
llama.cpp b9174 (vulkan)baselineagent4
75.0
59.0294.61.88s13.36185008.48s0.000 GiB
legacystack comparable
GeForce RTX 3090 · 24 GiBcap 200 Wdrv 590
llama.cpp 59778f0 (cuda)baselineagent4
64.0
52.2460.01.77s15.66185009.58s-0.010 GiB
legacystack comparable
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)unified
llama.cpp b8940 (vulkan)baselinechat1
55.3
52.9252.3148ms18.1371001.89s0.001 GiB
legacystack comparable
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)unified
llama.cpp b8940 (vulkan)baselinecodegen1
54.1
53.8457.4170ms18.571100018.58s0.000 GiB
legacystack comparable
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)unified
llama.cpp b8940 (vulkan)baselineagent1
53.7
51.61210.8446ms18.66185009.69s0.000 GiB
legacystack comparable
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)unified
llama.cpp b8940 (vulkan)baselinerag1
53.2
50.42234.5347ms18.810122003.97s0.000 GiB
legacystack comparable
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)unified
llama.cpp b8940 (vulkan)baselineagent4
35.1
25.5143.35.18s28.561850019.63s0.001 GiB

Environment

Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)
cpuAMD RYZEN AI MAX+ 395 w/ Radeon 8060S
gpuAMD Radeon 8060S
archStrix Halo (gfx1151)
vram96 GiB (system 31.1 GiB, unified)
hardware probes
copy 41% of theoryFP16 peak 30.3 TF
256-bit8000 MHz20 SM/CU
Microbenchmarks for memory copy and tensor math; raw-engine decode and API workload rows measure model-serving speed.
captheorycopyfp16bf16
fixed256 GB/s106 GB/s30.3 TF-
compute: 11.5
backendllama.cpp b8940 (cpu)
osUbuntu 24.04.4 LTS
kernel7.0.2-2-pve
python3.12.3
runs/cell3
warmups1
endpoint/v1/chat/completions
streamingtrue
GeForce RTX 3090 · 24 GiB
cpuAMD EPYC 7302P 16-Core Processor
gpuNVIDIA GeForce RTX 3090
archNVIDIA
vram24 GiB (system 64.0 GiB)
power350 W / 450 W max(78% cap)
pcieGen 4 x16 / Gen 4 x16 max
clocksgfx 1980/2100 MHz · mem 9501 MHz
temp43°C idle · 62°C peak
peak draw341 W
hardware probes
copy 42% of theoryFP16 peak 65.4 TFcopy/math flat across caps
384-bit9751 MHz82 SM/CU
Microbenchmarks for memory copy and tensor math; raw-engine decode and API workload rows measure model-serving speed.
captheorycopyfp16bf16
200 W936 GB/s391 GB/s65.4 TF65.4 TF
300 W936 GB/s391 GB/s65.4 TF65.3 TF
450 W936 GB/s391 GB/s65.4 TF65.4 TF
compute: 8.6
backendllama.cpp cuda-4f13cb7 (cuda)
osUbuntu 24.04 LTS
kernel6.17.13-7-pve
driverNVIDIA 590.48.01 + CUDA 13.1
libc2.39
python3.12.3
llama.cppversion: 18 (4f13cb7) built with GNU 13.3.0 for Linux x86_64
build flagsGGML_CUDA=ON CMAKE_BUILD_TYPE=Release
runs/cell5
warmups2
endpoint/v1/chat/completions
streamingtrue
GeForce RTX 3090 · 24 GiB
cpuAMD EPYC 7302P 16-Core Processor
gpuNVIDIA GeForce RTX 3090
archNVIDIA
vram24 GiB (system 64.0 GiB)
power450 W / 450 W max
pcieGen 4 x16 / Gen 4 x16 max
clocksgfx 1980/2100 MHz · mem 9501 MHz
temp53°C idle · 74°C peak
peak draw406 W
backendllama.cpp cuda-4f13cb7 (cuda)
osUbuntu 24.04 LTS
kernel6.17.13-7-pve
driverNVIDIA 590.48.01 + CUDA 13.1
libc2.39
python3.12.3
llama.cppversion: 18 (4f13cb7) built with GNU 13.3.0 for Linux x86_64
build flagsGGML_CUDA=ON CMAKE_BUILD_TYPE=Release
runs/cell5
warmups2
endpoint/v1/chat/completions
streamingtrue
GeForce RTX 3090 · 24 GiB
cpuAMD EPYC 7302P 16-Core Processor
gpuNVIDIA GeForce RTX 3090
archNVIDIA
vram24 GiB (system 64.0 GiB)
power200 W / 450 W max(44% cap)
backendllama.cpp 59778f0 (cuda)
osUbuntu 24.04 LTS
kernel6.17.13-7-pve
driver590.48.01
python3.12.3
runs/cell5
warmups2
endpoint/v1/chat/completions
streamingtrue
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)
cpuAMD RYZEN AI MAX+ 395 w/ Radeon 8060S
gpuAMD Radeon 8060S
archStrix Halo (gfx1151)
vram96 GiB (system 31.1 GiB, unified)
backendllama.cpp b1203 (rocm)
osUbuntu 24.04.4 LTS
kernel7.0.2-2-pve
python3.12.3
runs/cell3
warmups1
endpoint/v1/chat/completions
streamingtrue
GeForce RTX 5070 · 12 GiB
cpuAMD Ryzen 9 7900 12-Core Processor
gpuNVIDIA GeForce RTX 5070
archNVIDIA
vram11.94 GiB (system 30.4 GiB)
power250 W / 300 W max(83% cap)
hardware probes
copy 40% of theoryFP16 peak 69.6 TFcopy/math spread 2.5%
192-bit14001 MHz48 SM/CU
Microbenchmarks for memory copy and tensor math; raw-engine decode and API workload rows measure model-serving speed.
captheorycopyfp16bf16
200 W672 GB/s271 GB/s67.9 TF68.4 TF
250 W672 GB/s271 GB/s69.5 TF68.2 TF
300 W672 GB/s270 GB/s69.6 TF68.4 TF
compute: 12
backendllama.cpp b9174 (vulkan)
osCachyOS
kernel7.0.0-1-cachyos
driver595.58.03
python3.14.4
runs/cell5
warmups2
endpoint/v1/chat/completions
streamingtrue
Strix Halo · Radeon 8060S · 128 GiB unified (96 GiB VRAM)
cpuAMD RYZEN AI MAX+ 395 w/ Radeon 8060S
gpuAMD Radeon 8060S
archStrix Halo (gfx1151)
vram96 GiB (system 31.1 GiB, unified)
backendllama.cpp b8940 (vulkan)
osUbuntu 24.04.4 LTS
kernel7.0.2-2-pve
python3.12.3
runs/cell3
warmups1
endpoint/v1/chat/completions
streamingtrue