Llama
Llama 3.1 / 3.2
| Build (GGUF Q4_K_M) | Params | Status |
| Llama-3.2-3B-Instruct | 3B | Certified |
| Llama-3.2-1B-Instruct | 1B | Certified |
| Meta-Llama-3.1-8B-Instructreference model | 8B | Certified |
Qwen3, dense
includes abliterated
| Build (GGUF Q4_K_M) | Params | Status |
| Qwen3-8B | 8B | Certified |
| Qwen3-4B-Instruct-2507 | 4B | Certified |
| Qwen3-4B | 4B | Certified |
| Qwen3-14B | 14B | Beta |
| Qwen3-4B-Instruct-2507abliterated | 4B | Beta |
Qwen3-Next
hybrid linear-attention (Gated-DeltaNet)
| Build (GGUF Q4_K_M) | Params | Status |
| Ornith-1.0-9Bcommunity fine-tune | 9B | Certified |
| Qwen3.5-9B | 9B | Certified |
| Qwythos-9Bcommunity fine-tune | 9B | Certified |
Phi
Microsoft Phi-4
| Build (GGUF Q4_K_M) | Params | Status |
| Phi-4 | 14B | Certified |
| Phi-4-mini-instruct | 3.8B | Certified |
Mistral-Nemo
12B class
| Build (GGUF Q4_K_M) | Params | Status |
| Mistral-Nemo-12B-Instruct | 12B | Certified |
| Mistral-Nemo-Instruct-2407 | 12B | Certified |
Ministral
Ministral 3
| Build (GGUF Q4_K_M) | Params | Status |
| Ministral-3-3B-Instruct-2512 | 3B | Certified |
DeepSeek-R1 Distill
R1 reasoning distill on Qwen-14B
| Build (GGUF Q4_K_M) | Params | Status |
| DeepSeek-R1-Distill-Qwen-14B | 14B | Certified |
Gemma 3 / 4
gate-ready, steering in progress
| Build (GGUF Q4_K_M) | Params | Status |
| Gemma-4-E4B-it | 4Beffective | Beta |
| Gemma-3-4B-it | 4B | Beta |
| Gemma-3-12B-it | 12B | Certified |
| Gemma-4-12B-it | 12B | Certified |
SmolLM3
SmolLM3 3B
| Build (GGUF Q4_K_M) | Params | Status |
| SmolLM3-3B | 3B | Beta |
All builds are GGUF Q4_K_M. Steering artifacts (the per-model calibration
internals) are part of the licensed package and aren't published here. Grading is
ongoing as the fleet grows. Your policy packs are authored model-independently,
so they carry to any build unchanged, certified or beta.