Ayb 0.5 0.8B vs bison-1
Compare documented capabilities and access. A dash means a comparable figure is unavailable.
| Specification | ||
|---|---|---|
| Task | Armenian language and vision | Decision scoring |
| Published capabilities | Lightweight Armenian text processing and image understanding. Metric AI reports 53.00% on ArmBench, evaluated zero-shot. | Post-trained from Gemma 4 31B IT for classification, routing and rubric scoring. |
| Model size | 0.8B backbone tier | 31B parameters |
| Access | Downloadable weights after publisher approval on Hugging Face | Public weights and Baseten Model APIs as decision-models/baseten-bison-latest |
| Result | Text | Typed decisions and probabilities for supplied options |
| Operating limits | Adapted for Armenian and English, with training focused on Eastern Armenian. | Up to 131,072 tokens per question on Baseten Model APIs. |
| Billing terms | Not published | $0.10 USD per million input tokens on Baseten Model APIs. |
| Sources | Model weights · Additional documentation | Model weights · Additional documentation |
| Inputs | Text, Images | Text, Images |
| Context window | Not published | 131,072 tokens |
| Weights | Open weights · CC BY-NC 4.0 | Open weights · MIT and Apache-2.0 |
| Released | Nov 20, 2026 | Nov 20, 2026 |