This is an experimental preview of the CodeQ 9B model, created to verify the dataset parsing and tool schema during training.
It was trained for agentic coding and reasoning using one of the four CodeQ datasets. This model demonstrates reasoning and planning capabilities far superior to Qwen3.5-9B; it was trained on Ox-Alpha (GLM-5.3) traces and follows a "Research > Reason > Code > Verify" workflow for programming tasks.
This model is a test run for the final CodeQ-9B; please do not regard it as a finished product. Testing—and potentially benchmarking—is currently underway.
- Downloads last month
- 63
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support