Grouped by reported model identity. Exact weights, quantizations and software can differ. How results are reviewed
Documented results
24 selected tests| View Reportllm-bench.io | gemma-4-26B-A4B-it | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 69.85 / 100LLM judge · Mean | 810.64 tok/sPrefill · mean · per request | 46.46 tok/sGeneration · mean · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View ReportCustom evaluation | gemma-4-26B-A4B-itConflicting evidence | HumanEval100 problems | 96 %Reported pass@1 · n = 100 | — | 71.18 tok/sGeneration · mean · per request | 2 × W7800Layer split · 48 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | gemma-4-26B-A4B-it | MMLU57 subjects · first item per subject1 concurrent request | 77.19 %Accuracy · 44 / 57 · Exact-choice parser | — | — | 2 × R9700Tensor parallel · 32 GB VRAM/GPUConfiguration
|
| View ReportCoding session | gemma-4-26B-A4B-itPartial evaluation | SWE-bench Verified MiniVerified Mini: 50 tasks (25 Django + 25 Sphinx) | 50 %LLM judge · 25 / 50 52 %Exit-zero rate · 26 / 50 · Container test exit | — | — | 1 × R9700Single GPUConfiguration
|
| View ReportCustom evaluation | gemma-4-26B-A4B-it | GSM8KFull1319questions;5shot8K configured context | 95.83 %Accuracy · n = 100 · Original GSM8K tagged answer extraction and strict string comparison | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | 1,995.1 tok/sPrefill · unknown | 91.26 tok/sGeneration · unknown | 2 × RX 7900 XTXLayer split · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-26B-A4B-it | Token speed | — | 80.7 tok/sPrefill · per request | 17.3 tok/sGeneration · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | 1,831.96 tok/sPrefill · unknown | 77.32 tok/sGeneration · peak · unknown | 3 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | 721.53 tok/sPrefill · median · unknown | 45.08 tok/sGeneration · median · unknown | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportGeneric LLM | gemma-4-26B-A4B-it | Token speed | 24.2 tok/sThroughput · per request | — | — | 4 × R9700Tensor parallel · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed512 input tokens · 0 output tokens · 0 existing KV tokens | — | 909.04 tok/sPrefill · mean · per request | — | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed | — | 3,868 tok/sPrefill · per request | — | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed2,048 input tokens · 16,384 existing KV tokens | — | 2,147.45 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed | — | 3,079.41 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed | — | 3,427.14 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | gemma-4-26B-A4B-itConflicting evidence | Token speedsynth-bulk-dispatcher-impl | — | — | 69.9 tok/sGeneration · mean · per request | 2 × W7800Layer split · 48 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | gemma-4-26B-A4B-it | GSM8KFull1319questions;5shot8K configured context | 95.91 %Accuracy · n = 100 · Original GSM8K tagged answer extraction and strict string comparison | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | 1,823.93 tok/sPrefill · unknown | 107.99 tok/sGeneration · unknown | 2 × RX 7900 XTXLayer split · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | gemma-4-26B-A4B-it | Token speed | — | 61.3 tok/sPrefill · per request | 31 tok/sGeneration · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | 1,552.32 tok/sPrefill · unknown | 73.7 tok/sGeneration · peak · unknown | 3 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | gemma-4-26B-A4B-it | Token speed | — | — | 51.22 tok/sGeneration · unknown | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportGeneric LLM | gemma-4-26B-A4B-it | Token speed | — | — | 26 tok/sGeneration · per request | 4 × R9700Tensor parallel · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed0 input tokens · 128 output tokens · 0 existing KV tokens | — | — | 24.64 tok/sGeneration · mean · per request | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View Reportllama-bench | gemma-4-26B-A4B-it | Token speed | — | — | 157.2 tok/sGeneration · per request | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
Quantizations and software
Quantizations: BF16, FP8, INT4 AWQ (GPTQ-v2-fixed conversion), MXFP4_MOE, Q4_0, Q4_K_M, Q4_K_XL, Q8_0, Q8_K_XL, UD-Q4_K_M, UD-Q4_K_XL, UD-Q5_K_M, UD-Q6_K
Engines: llama.cpp, LM Studio, Ollama, SGLang, vLLM
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: W7800 · R9700 · RX 9070 · RX 7900 XTX · RX 7900 GRE · RX 7600 XT · Ryzen AI Max+ 395