Underdog-Saluki-27B-1.0-MTP-GGUF
Underdog Saluki 27B 1.0 with the Qwen3.8-27B multi-token-prediction (MTP) head added, so llama.cpp can run it with self-speculative decoding (--spec-type draft-mtp). The original Saluki release has no MTP head.
Params
27.0 B
Context
—
Downloads 30d
173 K
Likes
43
Download history
tracking started — chart appears after 7 days of snapshots (5 recorded)173 K downloads in the last 30 days
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| Underdog-Saluki-27B-1.0-IQ2-mix-MTP.gguf | Q2 | 8.3 GB | 13.7 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ ollama run underdog-saluki-27b-1-0-mtp-gguf # pin the quantization explicitly $ ollama run underdog-saluki-27b-1-0-mtp-gguf-q2
$ huggingface-cli download Kujira/Underdog-Saluki-27B-1.0-MTP-GGUF-GGUF \
Underdog-Saluki-27B-1.0-IQ2-mix-MTP.gguf --local-dir .
$ llama-cli -m Underdog-Saluki-27B-1.0-IQ2-mix-MTP.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/underdog-saluki-27b-1-0-mtp-gguf
{
"hf_id": "Kujira/Underdog-Saluki-27B-1.0-MTP-GGUF",
"params_b": 27.00,
"context_length": null,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 173411,
"vram_estimates": [
{ "quant": "Q2", "gb": 13.7 }
],
"updated_at": "2026-11-06T01:00:21Z"
}
Specifications
- Parameters
- 27.0 B
- Licence
- apache-2.0
- First seen on the Hub
- 2026-10-08
- Base model
- Underdog-Saluki-27B-1.0
- Training datasets
- undisclosed
- Added to our catalog
- 2026-11-06
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model