Qwen 2.5 (14B Instruct)
Qwen · ~14B parameters · 13B+ · Last reviewed 2026-03-20
Why we use it
Higher-capacity Qwen tier for private deployments that need extra headroom before distillation to sub-4B SLMs.
License summary
Apache 2.0 for common 14B instruct builds - confirm before merging into commercial distributions.
Typical deployment profiles
- VPC, multi-GPU
Focus tags
- General
- Multilingual
- Reasoning
Typical use cases
- Large VPC assistants
- Evaluation anchors
- RAG with long contexts