Views
No views yet




| Feature | Specification |
|---|---|
| Number of Parameters | 4 Billion |
| Input Modalities | Text, Image, Video |
| Context Length | 524,288 tokens (512K) natively |
| Training | Multi-Token Prediction (MTP), strong-to-weak distillation |
| Thinking Mode | Enabled (toggleable via enable_thinking parameter) |
| Benchmark | Score |
|---|---|
| MMLU-Pro | 74.2% |
| GPQA Diamond | 72.3% |
| HumanEval | 68.2% |
| LiveCodeBench | 51.4% |
| MATH | 45.8% |
1import neuralnode as nn
2
3MODEL_ID = "tokenaii/Horus-Hiero-Mini-4B-GGUF/Horus-Hiero-Mini-4B-Q6_K.gguf"
4DEVICE = "cpu" # Change to "cuda" for GPU acceleration
5
6# Download and load
7model = nn.HorusModel(MODEL_ID, device=DEVICE).load()
8
9# Use immediately
10response = model.chat([
11 {"role": "user", "content": "Translate this hieroglyph: 𓂋𓏏𓈖𓀀"}
12])
13print(response.content)1@misc{tokenai_horus_hiero_4b_mini,
2 title={Horus Hiero 4B Mini: Advanced Multilingual and Hieroglyphs Language Model},
3 author={Assem Sabry and TokenAI},
4 year={2026},
5 url={https://huggingface.co/tokenaii/Horus-Hiero-Mini-4B}
6}