Ternary-Bonsai-2-27B-Uncensored-Heretic-GGUF
A decensored version of Prism ML's Ternary-Bonsai-2-27B, with a Heretic LoRA baked into the ternary weights.
Params
27.0 B
Context
—
Downloads 30d
195 K
Likes
138
Download history
daily snapshots · 12 days219 K181 K
Oct 14Oct 18Oct 22Oct 25
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 |
|---|---|---|---|---|
| mmproj-Q8_0.gguf | Q8_0 | 0.6 GB | 5.2 GB | ✅ Runs comfortably |
| mmproj-BF16.gguf | GGUF | 0.9 GB | 5.6 GB | ✅ Runs comfortably |
| Ternary-Bonsai-2-27B-Uncensored-Heretic-PTQ1_0.gguf | Q1_0 | 5.9 GB | 11.1 GB | ✅ Runs comfortably |
| Ternary-Bonsai-2-27B-Uncensored-Heretic-PQ2_0.gguf | Q2_0 | 7.2 GB | 12.5 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 ternary-bonsai-2-27b-uncensored-heretic-gguf # pin the quantization explicitly $ ollama run ternary-bonsai-2-27b-uncensored-heretic-gguf-q2_0
$ huggingface-cli download OS-Software/Ternary-Bonsai-2-27B-Uncensored-Heretic-GGUF-GGUF \
Ternary-Bonsai-2-27B-Uncensored-Heretic-PQ2_0.gguf --local-dir .
$ llama-cli -m Ternary-Bonsai-2-27B-Uncensored-Heretic-PQ2_0.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/ternary-bonsai-2-27b-uncensored-heretic-gguf
{
"hf_id": "OS-Software/Ternary-Bonsai-2-27B-Uncensored-Heretic-GGUF",
"params_b": 27.00,
"context_length": null,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 194945,
"vram_estimates": [
{ "quant": "Q2_0", "gb": 12.5 },
{ "quant": "Q1_0", "gb": 11.1 }
],
"updated_at": "2026-10-14T01:00:22Z"
}
Specifications
- Parameters
- 27.0 B
- Licence
- apache-2.0
- First seen on the Hub
- 2026-09-18
- Base model
- Ternary-Bonsai-2-27B-gguf
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-14
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model