Model benchmarks
unsloth/gemma-4-12B-it-qat-GGUF local LLM performance
As of September 2026, unsloth/gemma-4-12B-it-qat-GGUF runs at up to 51.4 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
12B
Peak speed
51.4 tok/s
Average speed
50.7 tok/s
Avg PP
697.9 tok/s
Min memory
5.9 GB
Max context
32.768 tokens
Avg output / run
10.998 tokens
Avg runtime / run
3m 44s
Avg quality
68.7
Benchmark runs
2
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 2 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 | 67.9 | 68.7 | 69.6 |
| Agent Workflow | 70.9 | 74.1 | 77.2 |
| Code Generation | 44.8 | 52.7 | 60.5 |
| Role Play & Narrative | 81.2 | 81.8 | 82.3 |
| Research & Analysis | 65.7 | 66.5 | 67.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination unsloth/gemma-4-12B-it-qat-GGUF has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 4070 | llama.cpp | — | 51.4 tok/s | 50.7 tok/s | 5.9 GB | 32.768 tokens | 68.7 | 2 |
Benchmark runs
All 2 unsloth/gemma-4-12B-it-qat-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is unsloth/gemma-4-12B-it-qat-GGUF good for coding?
- In our benchmarks, unsloth/gemma-4-12B-it-qat-GGUF scores 52.7/100 for coding. It runs at about 50.7 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
- Is unsloth/gemma-4-12B-it-qat-GGUF good for agentic (tool-using) tasks?
- In our benchmarks, unsloth/gemma-4-12B-it-qat-GGUF scores 74.1/100 for agentic workflows. It runs at about 50.7 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is unsloth/gemma-4-12B-it-qat-GGUF for local inference?
- Across 2 community benchmark runs, unsloth/gemma-4-12B-it-qat-GGUF reaches up to 51.4 tok/s and averages 50.7 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
- How much memory does unsloth/gemma-4-12B-it-qat-GGUF need?
- The leanest observed configuration used about 5.9 GB of memory.
- Which tools have been used to run unsloth/gemma-4-12B-it-qat-GGUF?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.