Model benchmarks
Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL local LLM performance
As of October 2026, Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL runs at up to 24.6 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
31B
Peak speed
24.6 tok/s
Average speed
24.3 tok/s
Avg PP
561.2 tok/s
Min memory
31.5 GB
Max context
65,536 tokens
Avg output / run
9,472 tokens
Avg runtime / run
6m 45s
Avg quality
74.7
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 | 73.7 | 74.7 | 75.7 |
| Agent Workflow | 64.0 | 74.1 | 84.3 |
| Code Generation | 66.7 | 66.8 | 66.8 |
| Role Play & Narrative | 84.8 | 87.0 | 89.2 |
| Research & Analysis | 67.1 | 70.9 | 74.8 |
Performance by hardware and tool
Every hardware/tool/quantization combination Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL 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 | llama.cpp | Q6_K | 24.6 tok/s | 24.3 tok/s | 31.5 GB | 65,536 tokens | 74.7 | 2 |
Benchmark runs
All 2 Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL good for coding?
- In our benchmarks, Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL scores 66.8/100 for coding. It runs at about 24.3 tok/s, so if you want more speed, JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M is faster (~40.4 tok/s) and still scores well for coding (84.9/100).
- Is Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL good for agentic (tool-using) tasks?
- In our benchmarks, Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL scores 74.1/100 for agentic workflows. It runs at about 24.3 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 Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL for local inference?
- Across 2 community benchmark runs, Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL reaches up to 24.6 tok/s and averages 24.3 tok/s, with the fastest results on NVIDIA GeForce RTX 3080 Ti.
- How much memory does Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL need?
- The leanest observed configuration used about 31.5 GB of memory (quantizations tested: Q6_K).
- Which tools have been used to run Gryphe/Gemma-4-31B-StyleTune:MY-Q6_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.