Model benchmarks
peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL local LLM performance
As of October 2026, peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL runs at up to 157.1 tok/s for local inference (best of 2 community benchmark runs across 2 GPUs).
Model size
2B
Peak speed
157.1 tok/s
Average speed
109.2 tok/s
Avg PP
1373.6 tok/s
Min memory
5.5 GB
Max context
131,072 tokens
Avg output / run
14,200 tokens
Avg runtime / run
3m 8s
Avg quality
55.5
Benchmark runs
2
GPUs tested
2
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 | 54.6 | 55.5 | 56.3 |
| Agent Workflow | 75.7 | 76.8 | 77.8 |
| Code Generation | 15.0 | 23.5 | 32.1 |
| Role Play & Narrative | 41.3 | 47.7 | 54.1 |
| Research & Analysis | 73.6 | 73.9 | 74.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF: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 4070 | llama.cpp | Q6_K | 157.1 tok/s | 157.1 tok/s | 5.8 GB | 131,072 tokens | 54.5 | 1 |
| NVIDIA GeForce RTX 2060 | llama.cpp | Q6_K | 61.3 tok/s | 61.3 tok/s | 5.5 GB | 65,536 tokens | 56.4 | 1 |
Benchmark runs
All 2 peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL good for agentic (tool-using) tasks?
- In our benchmarks, peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL scores 76.8/100 for agentic workflows. It runs at about 109.2 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP is faster (~144.4 tok/s) and still scores well for agentic workflows (82.6/100).
- How fast is peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL for local inference?
- Across 2 community benchmark runs, peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL reaches up to 157.1 tok/s and averages 109.2 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
- How much memory does peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL need?
- The leanest observed configuration used about 5.5 GB of memory (quantizations tested: Q6_K).
- Which tools have been used to run peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.