Qwen3.8 9B Distill
9.7B parameter open-weight model. A full-parameter distillation of a Qwen3.8 2.4T/A95B-scale teacher into the Qwen3.5-9B architecture — a Gated DeltaNet hybrid, not plain GQA, so ordinary KV-cache arithmetic doesn't transfer. Native 262K context, function calling per the Qwen3.5 spec. No official Ollama library tag (registry checked 2026-08-21) — built via a from-source llama.cpp conversion; Q4_K_M measured 5.4 GB on disk, 95.15 tok/s generation on an RTX 3080 (2570server, 2026-08-21). Publisher's own lm-evaluation-harness figures show a large MMLU gain over the Qwen3.5-9B base (0.546 to 0.751 flexible-extract) alongside a small GSM8K regression (0.885 to 0.870).
Empero AI · Qwen 3.8 Distill
Editorial review
VRAM figures are empirical estimates. Actual usage varies by runtime, context length, and system configuration. Verify on your specific hardware before production use.
Will Qwen3.8 9B Distill run on your machine?
Qwen3.8 9B Distill is 9.7B parameters and needs 6.9 GB of VRAM at Q4_K_M — 5.4 GB of weights plus 1.5 GB of runtime overhead for the inference server itself.
VRAM by quantization
| Quantization | Weights | Needs (with overhead) | Quality |
|---|---|---|---|
| Q4_K_M | 5.4 GB | 6.9 GB | good |
Fit on common hardware at Q4_K_M
| Hardware | Memory the model can use | System RAM | Verdict |
|---|---|---|---|
| CPU Only | None (CPU only) | 16 GB | CPU offload |
| RTX 4060 Laptop | 8 GB | 16 GB | Tight |
| RTX 3060 (12GB) | 12 GB | 32 GB | Comfortable |
| RTX 4060 Ti (16GB) | 16 GB | 32 GB | Comfortable |
| RTX 3090 | 24 GB | 64 GB | Comfortable |
Comfortable means VRAM clears the requirement by 2 GB or more. Tight means it covers the requirement with no margin. CPU offload means the model does not fit in VRAM but system RAM is at least 1.6× the weights, so it will run at reduced speed — expect roughly 1–5 tokens per second. Figures are weights plus a fixed runtime overhead and exclude KV-cache growth, which scales with context length.
Need more hardware for Qwen3.8 9B Distill? Open the PC Builder for the 13B / 14B tier →
This catalog has no verified local Ollama tag for this checkpoint, so fit grades do not include a local run command.
Ready to run this model locally?
Find a compatible interface in our Local AI Tools directory →