Model benchmarks

gemma-4-26b-a4b-q4kxl local LLM performance

As of September 2026, gemma-4-26b-a4b-q4kxl runs at up to 171.1 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).

llama.cpp
ShareRedditX

Model size

26B

Peak speed

171.1 tok/s

Average speed

169.9 tok/s

Min memory

12.7 GB

Max context

65,536 tokens

Avg output / run

15,944 tokens

Avg runtime / run

1m 34s

Avg quality

69.6

Benchmark runs

3

GPUs tested

1

Quality by task

Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 3 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

TaskP5 (low)AvgP95 (high)
Overall65.469.673.7
Agent Workflow73.175.278.1
Code Generation64.167.973.3
Role Play & Narrative69.179.987.1
Research & Analysis46.255.365.3

Performance by hardware and tool

Every hardware/tool/quantization combination gemma-4-26b-a4b-q4kxl has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
AMD Radeon RX 7900 XT/7900 XTX/7900 GRE/7900Mllama.cpp171.1 tok/s169.9 tok/s12.7 GB65,536 tokens69.63

Frequently asked questions

Is gemma-4-26b-a4b-q4kxl good for coding?
In our benchmarks, gemma-4-26b-a4b-q4kxl scores 67.9/100 for coding. It runs at about 169.9 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for coding (66.9/100).
Is gemma-4-26b-a4b-q4kxl good for agentic (tool-using) tasks?
In our benchmarks, gemma-4-26b-a4b-q4kxl scores 75.2/100 for agentic workflows. It runs at about 169.9 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for agentic workflows (71.7/100).
How fast is gemma-4-26b-a4b-q4kxl for local inference?
Across 3 community benchmark runs, gemma-4-26b-a4b-q4kxl reaches up to 171.1 tok/s and averages 169.9 tok/s, with the fastest results on AMD Radeon RX 7900 XT/7900 XTX/7900 GRE/7900M.
How much memory does gemma-4-26b-a4b-q4kxl need?
The leanest observed configuration used about 12.7 GB of memory.
Which tools have been used to run gemma-4-26b-a4b-q4kxl?
Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.