Cloudflare · 27B · Dense
Decision model on Qwen 3.8 27B — calibrated answers over text, JSON, images and video, with a Jev-compatible API Check if your GPU or Mac can run Clef locally — 15.3 GB min, 25.5 GB recommended.
2026-09256K context
| Quant | Bits | VRAM | Quality | Status |
|---|---|---|---|---|
| Q2_K | 2 | 9.3 GB | low | — |
| Q3_K_M | 3 | 12.8 GB | moderate | — |
| Q4_K_M | 4 | 14.5 GB | good | — |
| Q5_K_M | 5 | 18 GB | good | — |
| Q6_K | 6 | 21.6 GB | excellent | — |
| Q8_0 | 8 | 28.6 GB | excellent | — |
| F16 | 16 | 56.6 GB | lossless | — |
Can I run Clef locally?
- Can I run Clef locally?
- Clef needs about 15.3 GB of memory at a minimum and 25.5 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does Clef need?
- At Q4_K_M, Clef uses about 14.5 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.