Model benchmarks
OrcaCyber-Nano-1.0-AI-Defense-Mix-13G local LLM performance
As of October 2026, OrcaCyber-Nano-1.0-AI-Defense-Mix-13G runs at up to 44.4 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
44.4 tok/s
Average speed
43.5 tok/s
Avg PP
323.2 tok/s
Min memory
16.9 GB
Max context
70,000 tokens
Avg output / run
28,447 tokens
Avg runtime / run
10m 33s
Avg quality
82.6
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 | 81.7 | 82.6 | 83.6 |
| Agent Workflow | 81.7 | 85.9 | 90.1 |
| Code Generation | 76.8 | 80.0 | 83.2 |
| Role Play & Narrative | 71.3 | 80.9 | 90.5 |
| Research & Analysis | 82.0 | 83.7 | 85.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination OrcaCyber-Nano-1.0-AI-Defense-Mix-13G 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 |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7800 XT | llama.cpp k8 vTurbo4 | — | 44.4 tok/s | 44.4 tok/s | 16.9 GB | 65,536 tokens | 83.7 | 1 |
| AMD Radeon RX 7800 XT | llama.cpp k4 v4 | — | 42.7 tok/s | 42.7 tok/s | 18.2 GB | 70,000 tokens | 81.6 | 1 |
Benchmark runs
All 2 OrcaCyber-Nano-1.0-AI-Defense-Mix-13G runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is OrcaCyber-Nano-1.0-AI-Defense-Mix-13G good for coding?
- In our benchmarks, OrcaCyber-Nano-1.0-AI-Defense-Mix-13G scores 80.0/100 for coding. It runs at about 43.5 tok/s, so if you want more speed, nvidia/Qwen3.8-27B-NVFP4 is faster (~75.5 tok/s) and still scores well for coding (83.5/100).
- Is OrcaCyber-Nano-1.0-AI-Defense-Mix-13G good for agentic (tool-using) tasks?
- In our benchmarks, OrcaCyber-Nano-1.0-AI-Defense-Mix-13G scores 85.9/100 for agentic workflows. It runs at about 43.5 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
- How fast is OrcaCyber-Nano-1.0-AI-Defense-Mix-13G for local inference?
- Across 2 community benchmark runs, OrcaCyber-Nano-1.0-AI-Defense-Mix-13G reaches up to 44.4 tok/s and averages 43.5 tok/s, with the fastest results on AMD Radeon RX 7800 XT.
- How much memory does OrcaCyber-Nano-1.0-AI-Defense-Mix-13G need?
- The leanest observed configuration used about 16.9 GB of memory.
- Which tools have been used to run OrcaCyber-Nano-1.0-AI-Defense-Mix-13G?
- Benchmarks were submitted using llama.cpp k4 v4, llama.cpp k8 vTurbo4. Results are community-contributed and updated as new runs arrive.