Ayb 0.5 0.8B vs Inkling 2 Mini
Compare documented capabilities and access. A dash means a comparable figure is unavailable.
| Specification | ||
|---|---|---|
| Task | Armenian language and vision | Multimodal reasoning and coding |
| Published capabilities | Lightweight Armenian text processing and image understanding. Metric AI reports 53.00% on ArmBench, evaluated zero-shot. | Reuses its middle layers three times per token. Thinking Machines reports 48.2% on DeepSWE v1.1. |
| Model size | 0.8B backbone tier | 26.4B total / 3.4B active parameters |
| Access | Downloadable weights after publisher approval on Hugging Face | Downloadable weights and Tinker inference and customization |
| Result | Text | Not published |
| Operating limits | Adapted for Armenian and English, with training focused on Eastern Armenian. | Native mixed-precision checkpoint occupies approximately 20 GB. |
| Sources | Model weights · Additional documentation | Model weights · Additional documentation |
| Inputs | Text, Images | Text, Images |
| Weights | Open weights · CC BY-NC 4.0 | Open weights · Apache-2.0 |
| Released | Nov 20, 2026 | Nov 20, 2026 |