Model benchmarks
cyankiwi/gemma-4-31B-it-AWQ-4bit local LLM performance
As of October 2026, cyankiwi/gemma-4-31B-it-AWQ-4bit runs at up to 32.7 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
31B
Peak speed
32.7 tok/s
Average speed
32.7 tok/s
Avg PP
852.4 tok/s
Min memory
n/a
Max context
16,384 tokens
Avg output / run
8,668 tokens
Avg runtime / run
4m 32s
Avg quality
74.8
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 | 71.8 | 74.8 | 77.7 |
| Agent Workflow | 81.1 | 83.4 | 85.7 |
| Code Generation | 65.3 | 68.6 | 71.8 |
| Role Play & Narrative | 71.9 | 76.0 | 80.0 |
| Research & Analysis | 64.4 | 71.1 | 77.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination cyankiwi/gemma-4-31B-it-AWQ-4bit 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3080 Ti | vLLM | — | 32.7 tok/s | 32.7 tok/s | n/a | 16,384 tokens | 74.8 | 2 |
Benchmark runs
All 2 cyankiwi/gemma-4-31B-it-AWQ-4bit runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is cyankiwi/gemma-4-31B-it-AWQ-4bit good for coding?
- In our benchmarks, cyankiwi/gemma-4-31B-it-AWQ-4bit scores 68.6/100 for coding. It runs at about 32.7 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is cyankiwi/gemma-4-31B-it-AWQ-4bit good for agentic (tool-using) tasks?
- In our benchmarks, cyankiwi/gemma-4-31B-it-AWQ-4bit scores 83.4/100 for agentic workflows. It runs at about 32.7 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 cyankiwi/gemma-4-31B-it-AWQ-4bit for local inference?
- Across 2 community benchmark runs, cyankiwi/gemma-4-31B-it-AWQ-4bit reaches up to 32.7 tok/s and averages 32.7 tok/s, with the fastest results on NVIDIA GeForce RTX 3080 Ti.
- Which tools have been used to run cyankiwi/gemma-4-31B-it-AWQ-4bit?
- Benchmarks were submitted using vLLM. Results are community-contributed and updated as new runs arrive.