Model benchmarks
Ternary-Bonsai-2-27B-PQ2_0 local LLM performance
As of October 2026, Ternary-Bonsai-2-27B-PQ2_0 runs at up to 44.5 tok/s for local inference (best of 4 community benchmark runs across 3 GPUs).
Model size
27B
Peak speed
44.5 tok/s
Average speed
32.9 tok/s
Avg PP
480.3 tok/s
Min memory
10.2 GB
Max context
131,072 tokens
Avg output / run
31,923 tokens
Avg runtime / run
19m 20s
Avg quality
60.2
Benchmark runs
4
GPUs tested
3
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 4 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 | 44.8 | 60.2 | 77.0 |
| Agent Workflow | 24.0 | 62.7 | 80.3 |
| Code Generation | 2.2 | 27.3 | 61.6 |
| Role Play & Narrative | 69.3 | 82.1 | 88.8 |
| Research & Analysis | 37.7 | 68.5 | 87.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination Ternary-Bonsai-2-27B-PQ2_0 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 4070 | prism | Q2_0 | 44.5 tok/s | 44.5 tok/s | 10.4 GB | 65,536 tokens | 68.2 | 1 |
| NVIDIA GeForce RTX 5060 | llama.cpp | Q2_0 | 41.8 tok/s | 41.6 tok/s | n/a | 131,072 tokens | 61.2 | 2 |
| AMD Radeon 680M | llama.cpp | Q2_0 | 4.0 tok/s | 4.0 tok/s | 10.2 GB | 8,192 tokens | 50.1 | 1 |
Benchmark runs
All 4 Ternary-Bonsai-2-27B-PQ2_0 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Ternary-Bonsai-2-27B-PQ2_0 good for coding?
- In our benchmarks, Ternary-Bonsai-2-27B-PQ2_0 scores 27.3/100 for coding. It runs at about 32.9 tok/s, so if you want more speed, nvidia/Qwen3.8-27B-NVFP4 is faster (~63.3 tok/s) and still scores well for coding (82.7/100).
- Is Ternary-Bonsai-2-27B-PQ2_0 good for agentic (tool-using) tasks?
- In our benchmarks, Ternary-Bonsai-2-27B-PQ2_0 scores 62.7/100 for agentic workflows. It runs at about 32.9 tok/s, so if you want more speed, swift-iq2_xs is faster (~91.5 tok/s) and still scores well for agentic workflows (91.4/100).
- How fast is Ternary-Bonsai-2-27B-PQ2_0 for local inference?
- Across 4 community benchmark runs, Ternary-Bonsai-2-27B-PQ2_0 reaches up to 44.5 tok/s and averages 32.9 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
- How much memory does Ternary-Bonsai-2-27B-PQ2_0 need?
- The leanest observed configuration used about 10.2 GB of memory (quantizations tested: Q2_0).
- Which tools have been used to run Ternary-Bonsai-2-27B-PQ2_0?
- Benchmarks were submitted using llama.cpp, prism. Results are community-contributed and updated as new runs arrive.