OmniJev/OneJev-9B
OmniJev/OneJev-9B is a 9 billion parameter multimodal System One decision model developed by OmniJev, fine-tuned from Qwen/Qwen3.5-9B. It processes screenshots, photos, videos, or plain text inputs to return calibrated probabilities for decision options in a single forward pass. This model excels at agentic decision-making tasks, offering rapid response times for complex queries, with a context length of 32768 tokens.
Loading preview...
OneJev-9B: A Multimodal System One Decision Model
OneJev-9B is a 9 billion parameter multimodal model developed by OmniJev, specifically designed for System One decision-making. It is a full fine-tune of Qwen/Qwen3.5-9B, trained on 99,193 questions derived from real agent runs, videos, and images. This model is engineered to take diverse inputs, including screenshots, photos, videos, or plain text, and provide calibrated probabilities for various options in a single forward pass.
Key Capabilities
- Multimodal Input Processing: Handles images (screenshots, photos), video, and text simultaneously.
- System One Decision Making: Provides rapid, calibrated probability outputs for decision options.
- High Efficiency: On an H200 GPU, OneJev-9B can answer 1 question in 81 ms or 10 questions in 131 ms (13.1 ms per question) with a 1280x720 screenshot.
- Agentic Task Optimization: Fine-tuned on real-world agent interactions for robust performance in automated decision scenarios.
Good For
- Automated Agent Systems: Ideal for applications requiring quick, probabilistic decisions based on visual and textual context.
- Real-time Interaction Analysis: Suitable for processing live screenshots or video frames to inform immediate actions.
- Complex Decision Support: Provides a structured approach to evaluating multiple options with calibrated probabilities.
- Benchmarking against other models: Demonstrates competitive accuracy against models like Jev 1.13 and Jev-Omni 12B on its specialized test set.