Model benchmarks
pq2 local LLM performance
As of September 2026, pq2 runs at up to 48.8 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
48.8 tok/s
Average speed
48.4 tok/s
Avg PP
595.9 tok/s
Min memory
3.4 GB
Max context
8.192 tokens
Avg output / run
15.936 tokens
Avg runtime / run
5m 38s
Avg quality
65.9
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 | 62.4 | 65.9 | 69.3 |
| Agent Workflow | 85.0 | 85.4 | 85.9 |
| Code Generation | 8.5 | 15.1 | 21.7 |
| Role Play & Narrative | 79.4 | 83.9 | 88.4 |
| Research & Analysis | 76.9 | 79.1 | 81.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination pq2 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 | — | 48.8 tok/s | 48.4 tok/s | 3.4 GB | 8.192 tokens | 65.9 | 2 |
Benchmark runs
All 2 pq2 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is pq2 good for coding?
- In our benchmarks, pq2 scores 15.1/100 for coding. It runs at about 48.4 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 pq2 good for agentic (tool-using) tasks?
- In our benchmarks, pq2 scores 85.4/100 for agentic workflows. It runs at about 48.4 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 pq2 for local inference?
- Across 2 community benchmark runs, pq2 reaches up to 48.8 tok/s and averages 48.4 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
- How much memory does pq2 need?
- The leanest observed configuration used about 3.4 GB of memory.
- Which tools have been used to run pq2?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.