OPEN
N-Q-26M-JEV decision model
N-Q-26M-JEV scores Russian and English decisions with small local weights. Check the 1110 update, truncation limits and commercial licence terms.
By Q Project · Hugging Face weight release: 10 Nov 2026
A small Russian/English decision model with a custom hybrid backbone and decision head. The 1110 checkpoint adds toxicity and profanity training. It scores supplied options, with reversed-order averaging enabled by default; that serving choice uses two orders.
Quick answers
What is N-Q-26M-JEV by Q Project?
N-Q-26M-JEV: A small Russian/English decision model with a custom hybrid backbone and decision head.
Who makes N-Q-26M-JEV?
N-Q-26M-JEV: Q Project
Facts
- Size
- The card lists 29.7M parameters plus 16.8M n-gram parameters, despite the 26M name
- Inputs
- Russian and English text; choice, yes/no and score questions
- Limits
- Reads the first 1,536 state tokens; training-family wording is recommended
- Access
- Transformers custom code; safetensors and QPACK files
- Licence
- Q Project Model License 1.0: free non-commercial use; commercial use, including internal use, needs a paid licence. Company evaluation is allowed for up to 90 days without production or customer use.
- Data terms
- The card identifies non-commercial ANLI data and toxicity data without a stated licence; third-party terms still apply
Claims and sources
- The maker reports 45.9 on the 231-task public JevBench v1.2 run, down from 46.3 for 1027. Its reported Russian/English toxicity accuracies are 96.1%/90.1%; harmless English uses of violent words remain failure cases. These task results do not establish general reasoning ability. vendor reportedsource
Links
Source: Card and weights → Source: Licence → Source: Runtime →
Related projects
No related projects are listed in the registry.