gemma-4-E4B on AMD GPUs
5 results · 3 sources · 18 measurements
Grouped by reported model identity. Exact weights, quantizations and software can differ. How results are reviewed
Documented results
5 selected tests | | | | | | |
|---|
View
ReportLocalMaxxing | gemma-4-E4BQ4_K_MModel ref. ↗ | Token speed | — | 4,441.78 tok/sPrefill · unknown | 137.03 tok/sGeneration · unknown | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration- Memory
- 32 GB VRAM/GPU
- OS
- Windows 11 Enterprise 10.0.26200
- Engine
- llama.cpp
- Backend
- vulkan
- Harness
- LocalMaxxing community speed test; exact client where reported
- Prompt / output
- 512 / 512 tokens
- Context
- 2048
- Concurrency
- Not reported
- Date
- 2026-05-26
Full record → |
|---|
View
ReportLocalMaxxing | gemma-4-E4BQ8Model ref. ↗ | Token speed | — | 2,118.5 tok/sPrefill · unknown | 52.73 tok/sGeneration · unknown | 1 × RX 9060 XTSingle GPU · 16 GB VRAM/GPUConfiguration- Memory
- 16 GB VRAM/GPU
- OS
- Windows 11
- Engine
- llama.cpp
- Backend
- vulkan
- Harness
- LocalMaxxing community speed test; exact client where reported
- Prompt / output
- 8192 / 128 tokens
- Context
- 8192
- Concurrency
- Not reported
- Date
- 2026-05-09
Full record → |
|---|
View
ReportLocalMaxxing | gemma-4-E4BQ4_K_MModel ref. ↗ | Token speed | — | 313.6 tok/sPrefill · unknown | 71.2 tok/sGeneration · unknown | 1 × RX 7700 XTSingle GPU · 12 GB VRAM/GPUConfiguration- Memory
- 12 GB VRAM/GPU
- OS
- Windows 11
- Engine
- vllm
- Backend
- Not reported
- Harness
- LocalMaxxing community speed test; exact client where reported
- Prompt / output
- 37 / 1024 tokens
- Context
- 2048
- Concurrency
- Not reported
- Date
- 2026-07-25
Full record → |
|---|
View
ReportLocalMaxxing | gemma-4-E4BQ8_0Model ref. ↗ | Token speed | — | 5,076.22 tok/sPrefill · unknown | 95.95 tok/sGeneration · unknown | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration- Memory
- 32 GB VRAM/GPU
- OS
- Windows 11 Enterprise 10.0.26200
- Engine
- llama.cpp
- Backend
- vulkan
- Harness
- LocalMaxxing community speed test; exact client where reported
- Prompt / output
- 512 / 512 tokens
- Context
- 2048
- Concurrency
- Not reported
- Date
- 2026-05-26
Full record → |
|---|
View
ReportLocalMaxxing | gemma-4-E4BQ4_K_MModel ref. ↗ | Token speed | — | 521.54 tok/sPrefill · unknown | 25.34 tok/sGeneration · unknown | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration- Memory
- 32 GB VRAM/GPU
- OS
- Windows
- Engine
- llama.cpp
- Backend
- cuda
- Harness
- LocalMaxxing community speed test; exact client where reported
- Prompt / output
- 512 / 128 tokens
- Context
- 2048
- Concurrency
- Not reported
- Date
- 2026-06-23
Full record → |
|---|
Quantizations and software
Quantizations: Q4_K_M, Q8, Q8_0
Engines: llama.cpp, vllm
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: R9700 · RX 9060 XT · RX 7700 XT