Instructions to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1",
	filename="ornith-1.0-35b-IQ4_XS-MTP-graft-headQ6.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Use Docker

docker model run hf.co/LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

LM Studio
Jan

vLLM

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Ollama
How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Ollama:
```
ollama run hf.co/LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
```

Unsloth Studio

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 to start chatting

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Docker Model Runner
How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Docker Model Runner:
```
docker model run hf.co/LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M
```

Lemonade

How to use LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1:Q4_K_M

Run and chat with the model

lemonade run user.Ornith-1.0-35B-GGUF-llamacpp-tp1-Q4_K_M

List all available models

lemonade list

LordNeel commited on 28 days ago

Commit

84f7d57

verified ·

1 Parent(s): 271b7f5

Add release audit and artifact manifest

Browse files

Files changed (3) hide show

ARTIFACT_MANIFEST.json +356 -0
README.md +2 -0
RELEASE_AUDIT.md +84 -0

ARTIFACT_MANIFEST.json ADDED Viewed

	@@ -0,0 +1,356 @@

+{
+  "audited_remote_head_before_manifest_upload": "271b7f58edad5522fc4d800d22f7ccd6ac0b8789",
+  "generated_at": "2026-06-27T23:34:12.582029+00:00",
+  "harness_bundle": [
+    {
+      "blob_id": "6c9743c1480bd2c9bf5308626ffa1bd20eb56830",
+      "bytes": 7346,
+      "path": "MODEL_INVESTIGATION_HARNESS.md",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "b526e5c40a476b8edb38591dd2d72ad818a087a2",
+      "bytes": 134,
+      "path": "requirements-lab.txt",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "21ddcf93dfaae55111281e925f0c9974be2efd7b",
+      "bytes": 4597,
+      "path": "configs/model_behavior_suite.yaml",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "63d4c96c5aaace66eff3a86ce72c379c948422bc",
+      "bytes": 2287,
+      "path": "configs/serving_matrix.yaml",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "6ed937004209b063a95372ece3422f013996f3fe",
+      "bytes": 3621,
+      "path": "configs/quantization_matrix.yaml",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "93aa261c54ddf6804d2f62bb2f52e4d3a1d4a668",
+      "bytes": 842,
+      "path": "configs/experiment_lanes.yaml",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "8fd4bbc2a67c029bf0b360d0901ec200292887e6",
+      "bytes": 3192,
+      "path": "configs/training_plan.yaml",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "7b1a137d8ffe3d79652141a04cd9477bef67d831",
+      "bytes": 4573,
+      "path": "scripts/run_model_investigation.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "9a1232550356afcbde16a6eaf54a6bdbf466947d",
+      "bytes": 16017,
+      "path": "scripts/audit_model.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "846f1500d0f2aceeb2c97388d3e4b68439742e18",
+      "bytes": 2200,
+      "path": "scripts/check_audit_gates.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "d723430849c22f53b2a4d91a5fea9a427819edf0",
+      "bytes": 5117,
+      "path": "scripts/summarize_investigation.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "700ec81ab96b0573afdac1ffbaa84a9f0f94b903",
+      "bytes": 16665,
+      "path": "scripts/probe_model_behavior.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "3b9acd1e52f8203a59f1d72369f148e2bc672304",
+      "bytes": 7877,
+      "path": "scripts/probe_openai_contract.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "a5d63caae222f5df761668c58d36a158fa73653f",
+      "bytes": 8169,
+      "path": "scripts/bench_openai.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "1fefc910e4b1ebf41ca54c332d404918583f97a6",
+      "bytes": 4379,
+      "path": "scripts/collect_next_token_logprobs.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "aaf11e0827f98fe413959c07296cf1021d4def71",
+      "bytes": 6926,
+      "path": "scripts/compare_kld_logprobs.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "6e674b480252381cd367b480095ffa0d4e65df3a",
+      "bytes": 5718,
+      "path": "scripts/check_mtp_support.py",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "a1df912ce4ac635b80d1927dbe595d42a4c7d685",
+      "bytes": 22567,
+      "path": "scripts/audit_gguf_vllm_materialization.py",
+      "planned_in_this_upload": false,
+      "published": true
+    }
+  ],
+  "key_benchmark_docs": [
+    {
+      "blob_id": "5d2afc62b5b47faeab64cb323a741b1dcbd03bbb",
+      "bytes": 15106,
+      "path": "benchmarks/llamacpp-quant-benchmarks.md",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "376892b1847c3b98e8bd00b690466970de9fc910",
+      "bytes": 3294,
+      "path": "benchmarks/kld-quant-vs-bf16-top64.md",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "df0ab1ae4916ec9426ce7b58452ef0a979944af4",
+      "bytes": 5566,
+      "path": "benchmarks/llamacpp-iq4-xs-mtp-graft-adaptive-profile.md",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "18e0e00637602514a0c8b9d26805fe84e80bc471",
+      "bytes": 5962,
+      "path": "benchmarks/sft-prod-20k-preflight.md",
+      "planned_in_this_upload": false,
+      "published": true
+    }
+  ],
+  "key_probe_docs": [
+    {
+      "blob_id": "470ec1394c86b38eedb2a97ae32530b5a3329f48",
+      "bytes": 28352,
+      "path": "probes/investigation-ornith-q4-live.json",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "a6fe7fccd66e2d29e1d02571c54fbc1f415cf1fd",
+      "bytes": 1304,
+      "path": "probes/investigation-ornith-q4-live.md",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "ac4882041cf88837beecde1ead7418fd5262b28c",
+      "bytes": 1252,
+      "path": "probes/sft-prod-20k-stopped.json",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "c0376b5f05405e52dc747a1017c1860b45229c34",
+      "bytes": 21512,
+      "path": "probes/eval-lora-behavior-checkpoint-8000-fixed.json",
+      "planned_in_this_upload": false,
+      "published": true
+    },
+    {
+      "blob_id": "041fb5bb56dc25cf23b634437948011ae77ac1da",
+      "bytes": 1049,
+      "path": "probes/eval-lora-behavior-checkpoint-8000-fixed.gate.log",
+      "planned_in_this_upload": false,
+      "published": true
+    }
+  ],
+  "known_remaining_gaps": [
+    "No release-candidate SFT/post-training adapter exists because the 20k run was stopped and checkpoint-8000 failed behavior gates.",
+    "The integrated MTP profile is recommended adaptively; always-on MTP is still not a saturated c16 throughput win.",
+    "vLLM GGUF remains marked unusable for this model family in the documented probe; llama.cpp is the supported GGUF backend."
+  ],
+  "mtp_status": {
+    "high_concurrency_policy": "adaptive throttle; disable MTP when saturated",
+    "integrated_graft_artifact": "ornith-1.0-35b-IQ4_XS-MTP-graft-headQ6.gguf",
+    "low_concurrency_speedup": "c1/128 AR 221.8 tok/s vs MTP 279.4 tok/s; adaptive c1/128 319.5 tok/s",
+    "reference_report": "benchmarks/llamacpp-iq4-xs-mtp-graft-adaptive-profile.md"
+  },
+  "planned_upload_paths": [
+    "ARTIFACT_MANIFEST.json",
+    "README.md",
+    "RELEASE_AUDIT.md"
+  ],
+  "published_gguf": [
+    {
+      "blob_id": "22f67bcceb94b8b5678374fbc318c64f4f09efd0",
+      "bytes": 19632837760,
+      "gib": 18.285,
+      "path": "ornith-1.0-35b-IQ4_XS-MTP-graft-headQ6.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "b0d902aa4d45a51c18a85eaaf9ccb7db14cc8caa",
+      "bytes": 18939311680,
+      "gib": 17.639,
+      "path": "ornith-1.0-35b-IQ4_XS.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "a501a18c15af6d01f87b0a6a1ad730038555b3b4",
+      "bytes": 16764763712,
+      "gib": 15.613,
+      "path": "ornith-1.0-35b-Q3_K_M.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "fe22dc3ac4939a48113ce5c419550fb4c627d994",
+      "bytes": 21166757760,
+      "gib": 19.713,
+      "path": "ornith-1.0-35b-Q4_K_M.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "668278d63d45032bd0a9b3531805d8b821c00d24",
+      "bytes": 24729130848,
+      "gib": 23.031,
+      "path": "ornith-1.0-35b-Q5_K_M.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "6b508985d071c95cdfdfd709f9fc6a27ce44a78a",
+      "bytes": 28514152288,
+      "gib": 26.556,
+      "path": "ornith-1.0-35b-Q6_K.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "b1faf816685e013ceb18974c06599324f86e8402",
+      "bytes": 36903138880,
+      "gib": 34.369,
+      "path": "ornith-1.0-35b-Q8_0.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "a5ffefa3aebc758648ad0b082d7ccac31dad1759",
+      "bytes": 1149036128,
+      "gib": 1.07,
+      "path": "ornith-1.0-35b-mtp-bootstrap-layer39-IQ4_XS.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "7940ecc1866a24c202b258364449f7b848bdbdee",
+      "bytes": 3623587040,
+      "gib": 3.375,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r10prefix1000-h1r2-h3r8fw-lr5e6-Q6_K.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "11a278ab710bc917eab13aa1a7f1acca7e518173",
+      "bytes": 3623586912,
+      "gib": 3.375,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r10prefix250-h1r2-h3r8fw-lr5e6-Q6_K.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "3b02c9eb201e0c7b926bbf2d5572c1849a484d9a",
+      "bytes": 3623586912,
+      "gib": 3.375,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r10prefix500-h1r2-h3r8fw-lr5e6-Q6_K.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "3a9215c952b2faac6aed5f4f637810fa595aa3a2",
+      "bytes": 2501241952,
+      "gib": 2.329,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-IQ4_XS.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "18e1b2da75f41e21e44c52c6ac17856f350ad996",
+      "bytes": 3251055712,
+      "gib": 3.028,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-Q5_K_M.gguf",
+      "published": true
+    },
+    {
+      "blob_id": "ba41d0ef43fd5be68300e563973fad9428ab6bab",
+      "bytes": 3623586912,
+      "gib": 3.375,
+      "path": "ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-Q6_K.gguf",
+      "published": true
+    }
+  ],
+  "quant_status": {
+    "kld_report": "benchmarks/kld-quant-vs-bf16-top64.md",
+    "lowest_memory_validated": "ornith-1.0-35b-Q3_K_M.gguf",
+    "middle_footprint_validated": "ornith-1.0-35b-IQ4_XS.gguf",
+    "speed_focused_validated": "ornith-1.0-35b-Q4_K_M.gguf"
+  },
+  "readme_reference_check": {
+    "context_gguf_refs_not_expected_in_repo": [
+      "ornith-1.0-35b-bf16.gguf"
+    ],
+    "missing_advertised_gguf_refs": [],
+    "missing_advertised_non_gguf_refs": [],
+    "missing_non_gguf_refs": []
+  },
+  "repo_id": "LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1",
+  "schema_version": 1,
+  "serving_status": {
+    "multi_gpu_serving": "out_of_scope",
+    "recommended_integrated_mtp_profile": "IQ4_XS-MTP-graft-headQ6 with LLAMA_SPEC_MAX_DRAFTING_SLOTS=1 and SPEC_DRAFT_N_MAX=2",
+    "recommended_speed_profile": "llama.cpp Q4_K_M with REASONING=off",
+    "scope": "single_gpu_tp1_only"
+  },
+  "sft_status": {
+    "adapter_released": false,
+    "checkpoint_9000_exists": false,
+    "hf_stop_record": "probes/sft-prod-20k-stopped.json",
+    "latest_durable_checkpoint": "artifacts/train/ornith-35b-coding-lora-prod-20k/checkpoint-8000",
+    "non_durable_last_step_seen_in_log": 8673,
+    "release_blocker": "checkpoint-8000 behavior gate failed 4 checks; no adapter should be published from this run",
+    "training_state": "stopped_by_user"
+  },
+  "summary": {
+    "readme_advertised_file_refs": 62,
+    "readme_advertised_gguf_refs": 14,
+    "readme_missing_advertised_gguf_refs": 0,
+    "readme_missing_advertised_non_gguf_refs": 0,
+    "readme_missing_non_gguf_refs": 0,
+    "readme_non_gguf_refs": 48,
+    "remote_file_count_before_manifest_upload": 197,
+    "status": "pass"
+  }
+}

README.md CHANGED Viewed

@@ -35,6 +35,8 @@ Validated on 2026-06-26 with one RTX PRO 6000 Blackwell Max-Q 96GB GPU, `tp=1`,
 - `ornith-1.0-35b-Q5_K_M.gguf`
 - `ornith-1.0-35b-Q6_K.gguf`
 - `ornith-1.0-35b-Q8_0.gguf`
 - `MODEL_INVESTIGATION_HARNESS.md`
 - `requirements-lab.txt`
 - `configs/model_behavior_suite.yaml`

 - `ornith-1.0-35b-Q5_K_M.gguf`
 - `ornith-1.0-35b-Q6_K.gguf`
 - `ornith-1.0-35b-Q8_0.gguf`
+- `ARTIFACT_MANIFEST.json`
+- `RELEASE_AUDIT.md`
 - `MODEL_INVESTIGATION_HARNESS.md`
 - `requirements-lab.txt`
 - `configs/model_behavior_suite.yaml`

RELEASE_AUDIT.md ADDED Viewed

	@@ -0,0 +1,84 @@

+# Release Audit
+Generated: 2026-06-27T23:34:12.582029+00:00
+- HF repo: `LordNeel/Ornith-1.0-35B-GGUF-llamacpp-tp1`
+- Audited remote head before this audit upload: `271b7f58edad5522fc4d800d22f7ccd6ac0b8789`
+- Remote files before this audit upload: 197
+- README advertised files: 62 checked
+- README advertised GGUF files: 14 checked, 0 missing
+- README advertised non-GGUF files: 48 checked, 0 missing after planned upload
+- README broader docs/config/script references: 48 checked, 0 missing after planned upload
+- Overall reference status: `pass`
+## Published GGUF Artifacts
+| File | Published | GiB | Bytes |
+|---|---:|---:|---:|
+| `ornith-1.0-35b-IQ4_XS-MTP-graft-headQ6.gguf` | yes | 18.285 | 19632837760 |
+| `ornith-1.0-35b-IQ4_XS.gguf` | yes | 17.639 | 18939311680 |
+| `ornith-1.0-35b-Q3_K_M.gguf` | yes | 15.613 | 16764763712 |
+| `ornith-1.0-35b-Q4_K_M.gguf` | yes | 19.713 | 21166757760 |
+| `ornith-1.0-35b-Q5_K_M.gguf` | yes | 23.031 | 24729130848 |
+| `ornith-1.0-35b-Q6_K.gguf` | yes | 26.556 | 28514152288 |
+| `ornith-1.0-35b-Q8_0.gguf` | yes | 34.369 | 36903138880 |
+| `ornith-1.0-35b-mtp-bootstrap-layer39-IQ4_XS.gguf` | yes | 1.07 | 1149036128 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r10prefix1000-h1r2-h3r8fw-lr5e6-Q6_K.gguf` | yes | 3.375 | 3623587040 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r10prefix250-h1r2-h3r8fw-lr5e6-Q6_K.gguf` | yes | 3.375 | 3623586912 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r10prefix500-h1r2-h3r8fw-lr5e6-Q6_K.gguf` | yes | 3.375 | 3623586912 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-IQ4_XS.gguf` | yes | 2.329 | 2501241952 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-Q5_K_M.gguf` | yes | 3.028 | 3251055712 |
+| `ornith-1.0-35b-mtp-chaincorr-h2r9fw125-h1r2-h3r6ce375-lr5e7-Q6_K.gguf` | yes | 3.375 | 3623586912 |
+Context-only GGUF mentions not expected to exist in this repo, such as upstream/source inputs:
+- `ornith-1.0-35b-bf16.gguf`
+## Current Recommendations
+- Speed-focused serving: `Q4_K_M` through llama.cpp with `REASONING=off`.
+- Lowest-memory validated GGUF: `Q3_K_M`.
+- Middle-footprint validated GGUF: `IQ4_XS`.
+- Integrated MTP artifact: `ornith-1.0-35b-IQ4_XS-MTP-graft-headQ6.gguf`.
+- MTP serving policy: adaptive throttle with `LLAMA_SPEC_MAX_DRAFTING_SLOTS=1`; always-on MTP is not recommended for saturated c16 batches.
+- Multi-GPU serving remains out of scope; all serving profiles are single-GPU `tp=1`.
+## Harness Status
+| Path | Published | Planned in audit upload | Bytes |
+|---|---:|---:|---:|
+| `MODEL_INVESTIGATION_HARNESS.md` | yes | no | 7346 |
+| `requirements-lab.txt` | yes | no | 134 |
+| `configs/model_behavior_suite.yaml` | yes | no | 4597 |
+| `configs/serving_matrix.yaml` | yes | no | 2287 |
+| `configs/quantization_matrix.yaml` | yes | no | 3621 |
+| `configs/experiment_lanes.yaml` | yes | no | 842 |
+| `configs/training_plan.yaml` | yes | no | 3192 |
+| `scripts/run_model_investigation.py` | yes | no | 4573 |
+| `scripts/audit_model.py` | yes | no | 16017 |
+| `scripts/check_audit_gates.py` | yes | no | 2200 |
+| `scripts/summarize_investigation.py` | yes | no | 5117 |
+| `scripts/probe_model_behavior.py` | yes | no | 16665 |
+| `scripts/probe_openai_contract.py` | yes | no | 7877 |
+| `scripts/bench_openai.py` | yes | no | 8169 |
+| `scripts/collect_next_token_logprobs.py` | yes | no | 4379 |
+| `scripts/compare_kld_logprobs.py` | yes | no | 6926 |
+| `scripts/check_mtp_support.py` | yes | no | 5718 |
+| `scripts/audit_gguf_vllm_materialization.py` | yes | no | 22567 |
+## SFT Status
+- The production LoRA SFT run was stopped at user request.
+- Latest durable checkpoint: `checkpoint-8000`.
+- Training log reached step `8673`, but no `checkpoint-9000` was written.
+- No SFT adapter is released from this run.
+- Release blocker: `checkpoint-8000` failed the local behavior gate with 4 failed checks.
+## Remaining Gaps
+- No release-candidate SFT/post-training adapter exists because the 20k run was stopped and checkpoint-8000 failed behavior gates.
+- The integrated MTP profile is recommended adaptively; always-on MTP is still not a saturated c16 throughput win.
+- vLLM GGUF remains marked unusable for this model family in the documented probe; llama.cpp is the supported GGUF backend.
+## Machine-Readable Manifest
+See `ARTIFACT_MANIFEST.json`.