Is gemma4:26b-a4b-it-qat good for coding?

In our benchmarks, gemma4:26b-a4b-it-qat scores 61.1/100 for coding. It runs at about 68.1 tok/s, so if you want more speed, Nex-N2.5-mini-IQ3_XXS is faster (~91.4 tok/s) and still scores well for coding (81.6/100).

Is gemma4:26b-a4b-it-qat good for agentic (tool-using) tasks?

In our benchmarks, gemma4:26b-a4b-it-qat scores 75.0/100 for agentic workflows. It runs at about 68.1 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).

How fast is gemma4:26b-a4b-it-qat for local inference?

Across 17 community benchmark runs, gemma4:26b-a4b-it-qat reaches up to 121.4 tok/s and averages 68.1 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.

How much memory does gemma4:26b-a4b-it-qat need?

The leanest observed configuration used about 14.0 GB of memory (quantizations tested: Q4_0).

Which tools have been used to run gemma4:26b-a4b-it-qat?

Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.

Model benchmarks

gemma4:26b-a4b-it-qat local LLM performance

As of October 2026, gemma4:26b-a4b-it-qat runs at up to 121.4 tok/s for local inference (best of 17 community benchmark runs across 4 GPUs).

OllamaQ4_0

ShareReddit X

Model size

25.2B

Peak speed

121.4 tok/s

Average speed

68.1 tok/s

Avg PP

1651.7 tok/s

Min memory

14.0 GB

Max context

524,288 tokens

Avg output / run

17,629 tokens

Avg runtime / run

4m 17s

Avg quality

68.8

Benchmark runs

GPUs tested

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 17 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

Task	P5 (low)	Avg	P95 (high)
Overall	60.1	68.8	73.6
Agent Workflow	52.6	75.0	87.1
Code Generation	54.8	61.1	68.9
Role Play & Narrative	44.8	73.6	86.8
Research & Analysis	56.5	65.3	71.2

Performance by hardware and tool

Every hardware/tool/quantization combination gemma4:26b-a4b-it-qat has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
AMD Radeon RX 7900 XTX	Ollama	Q4_0	121.4 tok/s	116.0 tok/s	14.1 GB	131,072 tokens	69.2	3
CPU only	Ollama	Q4_0	78.6 tok/s	67.9 tok/s	14.0 GB	131,072 tokens	67.3	4
AMD Radeon RX 7800 XT	Ollama	Q4_0	73.3 tok/s	62.4 tok/s	14.2 GB	65,536 tokens	71.0	4
Apple M4 Pro	Ollama	Q4_0	62.3 tok/s	53.9 tok/s	14.0 GB	524,288 tokens	71.5	5
NVIDIA GeForce RTX 4070 Laptop GPU	Ollama	Q4_0	19.3 tok/s	19.3 tok/s	14.1 GB	8,192 tokens	51.3	1

Benchmark runs

All 17 gemma4:26b-a4b-it-qat runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is gemma4:26b-a4b-it-qat good for coding?: In our benchmarks, gemma4:26b-a4b-it-qat scores 61.1/100 for coding. It runs at about 68.1 tok/s, so if you want more speed, Nex-N2.5-mini-IQ3_XXS is faster (~91.4 tok/s) and still scores well for coding (81.6/100).
Is gemma4:26b-a4b-it-qat good for agentic (tool-using) tasks?: In our benchmarks, gemma4:26b-a4b-it-qat scores 75.0/100 for agentic workflows. It runs at about 68.1 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
How fast is gemma4:26b-a4b-it-qat for local inference?: Across 17 community benchmark runs, gemma4:26b-a4b-it-qat reaches up to 121.4 tok/s and averages 68.1 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
How much memory does gemma4:26b-a4b-it-qat need?: The leanest observed configuration used about 14.0 GB of memory (quantizations tested: Q4_0).
Which tools have been used to run gemma4:26b-a4b-it-qat?: Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.