Model
A security engineer's model. Runs locally.
OrcaCyber Nano 1.0 is a 27B vision-language reasoning model built for defensive security work — and packed as GGUF so it runs on your own hardware, offline, where the binaries and logs you are analyzing never leave the machine.
Built for: reverse engineering · vulnerability analysis · secure code review · triage · incident response
Security work is exactly the kind of work you do not want to send to a cloud endpoint: proprietary binaries, customer logs, internal source, unreleased vulnerabilities. GGUF + llama.cpp means the model runs on the same box as your tools, air-gapped if you need it.
It identifies as OrcaCyber Nano 1.0 AI Defense Model and keeps full general-purpose reasoning, so it is also a strong local coding and reasoning assistant.
| Benchmark | Score |
|---|---|
| MATH-500 | 98.60% |
| GSM8K | 97.50% |
| AIME 2025 | 96.23% |
| HumanEval+ | 93.29% |
| LiveCodeBench | 92.23% |
| GPQA Diamond | 88.28% |
| MBPP+ | 84.66% |
| IFEval | 82.81% |
| SWE-bench | 78.4% |
| Terminal-Bench 2.1 | 72.5% |
| IFBench | 56.33% |
| BigCodeBench | 49.82% |
MATH-500 █████████████████████████████████████████████████░ 98.60%
GSM8K █████████████████████████████████████████████████░ 97.50%
AIME 2025 ████████████████████████████████████████████████░░ 96.23%
HumanEval+ ███████████████████████████████████████████████░░░ 93.29%
LiveCodeBench ██████████████████████████████████████████████░░░░ 92.23%
GPQA Diamond ████████████████████████████████████████████░░░░░░ 88.28%
MBPP+ ██████████████████████████████████████████░░░░░░░░ 84.66%
IFEval █████████████████████████████████████████░░░░░░░░░ 82.81%
SWE-bench ███████████████████████████████████████░░░░░░░░░░░ 78.4%
Terminal-Bench 2.1 ████████████████████████████████████░░░░░░░░░░░░░░ 72.5%
IFBench ████████████████████████████░░░░░░░░░░░░░░░░░░░░░░ 56.33%
BigCodeBench █████████████████████████░░░░░░░░░░░░░░░░░░░░░░░░░ 49.82%
Benchmark harness, settings and full result tables are being finalized and will be expanded.
| File | Size | Contents |
|---|---|---|
OrcaCyber-Nano-1.0-AI-Defense-Mix-15.7G.gguf | 15.70 GB | best-quality build, with MTP head |
OrcaCyber-Nano-1.0-AI-Defense-Mix-13G.gguf | 12.99 GB | fits a 16 GB card with the MTP drafter |
OrcaCyber-Nano-1.0-AI-Defense-DFlash2-Q8_0.gguf | 2.06 GB | DFlash2 speculative drafter |
mmproj-OrcaCyber-Nano-1.0-AI-Defense-F16.gguf | 0.93 GB | vision projector (for image input) |
Loads in stock llama.cpp — no custom kernels.
hf download orcarouter/OrcaCyber-Nano-1.0 --local-dir ./ocn-gguf
# 16 GB card: Mix-13G + the built-in MTP drafter, 32k context
llama-server -m ocn-gguf/OrcaCyber-Nano-1.0-AI-Defense-Mix-13G.gguf \
--spec-type draft-mtp -np 1 -ngl 99 -c 32768
# 24 GB+ card: Mix-15.7G + DFlash2 drafter + vision
llama-server -m ocn-gguf/OrcaCyber-Nano-1.0-AI-Defense-Mix-15.7G.gguf \
--mmproj ocn-gguf/mmproj-OrcaCyber-Nano-1.0-AI-Defense-F16.gguf \
-md ocn-gguf/OrcaCyber-Nano-1.0-AI-Defense-DFlash2-Q8_0.gguf --spec-type draft-dflash \
--spec-draft-n-max 7 -ngl 99 -ngld 99 -c 32768
The MTP head is built into both main files; --spec-type draft-mtp turns it on with no extra file.
Pass the mmproj file for image input. llama-server defaults to 4 slots; pass -np 1 on small cards.
temperature = 1.0
top_p = 0.95
top_k = 20
| Architecture | Qwen3_5ForConditionalGeneration, VL hybrid |
| Layers / hidden | 64 / 5120 — 48 linear-attention + 16 full-attention |
| Context | 262,144 tokens |
| Modalities | text + image/video in, text out |
| Speculative decoding | MTP head (built in) + DFlash2 draft (separate file) |