Benchmark search

Try these searches

Found 39 run configurations

Each row is a run configuration (model + hardware + framework + quantization); repeated runs of the same setup are merged — raw counts appear in each result’s source line.

Modelgemma-4-26B-A4B-it

gemma-4-26B-A4B-it · IQ2_XS

L2 Cross-framework verified

AMD Ryzen AI Max+ 395 · 128GB · llama.cpp

Other

2026-10-15

92.57 tok/s

Decode

496.15 tok/s

Prefill

1.77 s

TTFT

28.31 GB

VRAM

Config (latest run)
Windows 11 23H2 CUDA N/A IQ2_XS llama.cpp v0.21 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · IQ3_XS

L2 Cross-framework verified

AMD Ryzen AI Max+ 395 · 128GB · llama.cpp

Other

2026-10-15

81.12 tok/s

Decode

341.45 tok/s

Prefill

1.66 s

TTFT

35.95 GB

VRAM

Config (latest run)
Windows 11 23H2 CUDA N/A IQ3_XS llama.cpp v0.40 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · IQ3_XS

L2 Cross-framework verified

AMD Ryzen AI Max+ 395 · 128GB · Ollama

Other

2026-10-15

79.36 tok/s

Decode

408.93 tok/s

Prefill

2.15 s

TTFT

35.95 GB

VRAM

Config (latest run)
Windows 11 23H2 CUDA N/A IQ3_XS Ollama v0.33 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · IQ3_XS

L2 Cross-framework verified

Apple M4 Pro · 64GB · vLLM

Other

2026-10-15

87.63 tok/s

Decode

428.03 tok/s

Prefill

0.53 s

TTFT

35.95 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A IQ3_XS vLLM v0.21 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q4_K_S

L2 Cross-framework verified

Apple M4 Pro · 64GB · vLLM

Other

2026-10-15

72.54 tok/s

Decode

364.75 tok/s

Prefill

1.6 s

TTFT

44.69 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q4_K_S vLLM v0.21 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q5_K_M

L2 Cross-framework verified

Apple M4 Pro · 64GB · Ollama

Other

2026-10-15

56.24 tok/s

Decode

299.52 tok/s

Prefill

0.72 s

TTFT

52.88 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q5_K_M Ollama v0.31

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q5_K_S

L2 Cross-framework verified

Apple M4 Pro · 64GB · Ollama

Other

2026-10-15

59.26 tok/s

Decode

228.14 tok/s

Prefill

0.61 s

TTFT

51.24 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q5_K_S Ollama v0.45 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q8_0

L2 Cross-framework verified

Apple M5 Pro · 64GB · llama.cpp

Other

2026-10-15

55.26 tok/s

Decode

261.28 tok/s

Prefill

1.58 s

TTFT

60.8 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q8_0 llama.cpp v0.37 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q6_K

L2 Cross-framework verified

Apple M5 Pro · 64GB · Strata

Other

2026-10-15

98.9 tok/s

Decode

379.8 tok/s

Prefill

1.66 s

TTFT

56.7 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q6_K Strata v0.52 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q3_K_S

L2 Cross-framework verified

Apple M5 Max · 128GB · vLLM

Other

2026-10-15

174.38 tok/s

Decode

573.36 tok/s

Prefill

1.98 s

TTFT

37.59 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q3_K_S vLLM v0.25 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q4_K_S

L2 Cross-framework verified

Apple M5 Max · 128GB · vLLM

Other

2026-10-15

136.41 tok/s

Decode

547.6 tok/s

Prefill

2.29 s

TTFT

44.69 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q4_K_S vLLM v0.37

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q2_K

L2 Cross-framework verified

Apple M5 Max · 128GB · Strata

Other

2026-10-15

228.98 tok/s

Decode

895.33 tok/s

Prefill

1.32 s

TTFT

32.13 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q2_K Strata v0.35 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q8_0

L2 Cross-framework verified

Apple M5 Max · 128GB · Strata

Other

2026-10-15

122.58 tok/s

Decode

429.76 tok/s

Prefill

1.24 s

TTFT

64.89 GB

VRAM

Config (latest run)
macOS 15.0 CUDA N/A Q8_0 Strata v0.53

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · IQ3_S

L2 Cross-framework verified

NVIDIA DGX Spark · 128GB · llama.cpp

Other

2026-10-15

112.41 tok/s

Decode

545.82 tok/s

Prefill

1.36 s

TTFT

38.14 GB

VRAM

Config (latest run)
Ubuntu 24.04 CUDA 12.8 IQ3_S llama.cpp v0.35

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q3_K_L

L2 Cross-framework verified

NVIDIA DGX Spark · 128GB · llama.cpp

Other

2026-10-15

105.58 tok/s

Decode

348.6 tok/s

Prefill

2.09 s

TTFT

41.41 GB

VRAM

Config (latest run)
Ubuntu 24.04 CUDA 12.8 Q3_K_L llama.cpp v0.23

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q3_K_L

L2 Cross-framework verified

NVIDIA DGX Spark · 128GB · Ollama

Other

2026-10-15

104.18 tok/s

Decode

516.57 tok/s

Prefill

2.18 s

TTFT

41.41 GB

VRAM

Config (latest run)
Ubuntu 24.04 CUDA 12.8 Q3_K_L Ollama v0.54 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q3_K_M

L2 Cross-framework verified

NVIDIA DGX Spark · 128GB · Ollama

Other

2026-10-15

113.11 tok/s

Decode

567.16 tok/s

Prefill

1.59 s

TTFT

39.23 GB

VRAM

Config (latest run)
Ubuntu 24.04 CUDA 12.8 Q3_K_M Ollama v0.35 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · IQ4_XS

L2 Cross-framework verified

NVIDIA Jetson AGX Orin 64GB · 64GB · vLLM

Other

2026-10-15

42.78 tok/s

Decode

174.63 tok/s

Prefill

2.05 s

TTFT

42.5 GB

VRAM

Config (latest run)
JetPack 6.0 CUDA 12.6 IQ4_XS vLLM v0.54 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q5_K_S

L2 Cross-framework verified

NVIDIA Jetson AGX Orin 64GB · 64GB · vLLM

Other

2026-10-15

34.07 tok/s

Decode

117.53 tok/s

Prefill

0.22 s

TTFT

51.24 GB

VRAM

Config (latest run)
JetPack 6.0 CUDA 12.6 Q5_K_S vLLM v0.46 Flash Attention

Sources: Other · 1 runs · 1 independent sources

gemma-4-26B-A4B-it · Q8_0

L2 Cross-framework verified

NVIDIA Jetson AGX Orin 64GB · 64GB · Ollama

Other

2026-10-15

24.68 tok/s

Decode

116.33 tok/s

Prefill

0.92 s

TTFT

60.8 GB

VRAM

Config (latest run)
JetPack 6.0 CUDA 12.6 Q8_0 Ollama v0.26 Flash Attention

Sources: Other · 1 runs · 1 independent sources

39 results

Related models
Related hardware
Related frameworks
Related quantizations