Instructions to use FermionResearch/Neutrino-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FermionResearch/Neutrino-8B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="FermionResearch/Neutrino-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("FermionResearch/Neutrino-8B", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B
# Run inference directly in the terminal:
llama cli -hf FermionResearch/Neutrino-8B

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf FermionResearch/Neutrino-8B
# Run inference directly in the terminal:
llama cli -hf FermionResearch/Neutrino-8B

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 FermionResearch/Neutrino-8B
# Run inference directly in the terminal:
./llama-cli -hf FermionResearch/Neutrino-8B

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 FermionResearch/Neutrino-8B
# Run inference directly in the terminal:
./build/bin/llama-cli -hf FermionResearch/Neutrino-8B

Use Docker

docker model run hf.co/FermionResearch/Neutrino-8B

LM Studio
Jan

vLLM

How to use FermionResearch/Neutrino-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FermionResearch/Neutrino-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FermionResearch/Neutrino-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FermionResearch/Neutrino-8B

SGLang

How to use FermionResearch/Neutrino-8B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "FermionResearch/Neutrino-8B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FermionResearch/Neutrino-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "FermionResearch/Neutrino-8B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FermionResearch/Neutrino-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use FermionResearch/Neutrino-8B with Ollama:
```
ollama run hf.co/FermionResearch/Neutrino-8B
```

Unsloth Studio

How to use FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for FermionResearch/Neutrino-8B to start chatting

How to use FermionResearch/Neutrino-8B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FermionResearch/Neutrino-8B

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": "FermionResearch/Neutrino-8B"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use FermionResearch/Neutrino-8B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FermionResearch/Neutrino-8B

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 FermionResearch/Neutrino-8B

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use FermionResearch/Neutrino-8B with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf FermionResearch/Neutrino-8B

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 "FermionResearch/Neutrino-8B" \
  --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 FermionResearch/Neutrino-8B with Docker Model Runner:
```
docker model run hf.co/FermionResearch/Neutrino-8B
```

Lemonade

How to use FermionResearch/Neutrino-8B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull FermionResearch/Neutrino-8B

Run and chat with the model

lemonade run user.Neutrino-8B-{{QUANT_TAG}}

List all available models

lemonade list

Neutrino-8B / MANIFEST.json

FermionResearch

runners: strict id-parser rebuild (B8) + native --draft everywhere + bin docs: MANIFEST.json

a661963 verified about 10 hours ago

Raw

History Blame Contribute Delete

16.4 kB

	{
	"note": "provenance: HF-tree filename -> source + sha256; weights entry is declarative until materialized",
	"weights_container": {
	"file": "neutrino-8b_v4.bin",
	"status": "FINAL \u2014 SHIP BRAIN (promoted 2026-07-27: GSM8K stated 49.33 / flexible 51.00 / BFCL 65.31 / IFEval-ps 73.17 / MMLU-Redux 67.84 / C4 21.475 / term 0.5067)",
	"bytes": 3875404812,
	"sha256": "016c6f362ad237a8cf4043601efbc1e766574f76ea58691e78fdef51b4155fa0",
	"source": "Fermion Research release archive (public copy: huggingface.co/fermionresearch/Neutrino-8B)",
	"coded_transport": {
	"file": "neutrino-8b_v4.tv4z",
	"bytes": 2559836297,
	"sha256": "83ec7a52bd6733e66fb546e89b9c6aa8b3411af7f648e9c0b2bb56ba50aac4df",
	"note": "lossless tv4z transport, 66.05% of raw; decode round-trip md5/sha/cmp PASS"
	}
	},
	"files": {
	".gitattributes": {
	"bytes": 1905,
	"sha256": "501edc5292358aaa11dd13cd5e9b25c43f94df3cc9fff43dd3c0df50efcc40c0",
	"source": "Fermion Research release archive"
	},
	"LICENSE": {
	"bytes": 11347,
	"sha256": "2f978a54ef89f3e4ee2d413c347e4b03500c95f29342277a287acc2ae27ae280",
	"source": "Apache-2.0 text (same file across all release trees)"
	},
	"NOTICE": {
	"bytes": 1557,
	"sha256": "a17bda9d5b1e7d90927a39a2b7c2e40a5cb8afe36ecb9f8c776802cec75b04ca",
	"source": "Fermion Research release archive"
	},
	"bin/MANIFEST.json": {
	"bytes": 1547,
	"sha256": "1aefbc115472a261545e0c6bb8a13636153de0a92514534a228e5e7583421799",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"bin/README.md": {
	"bytes": 12057,
	"sha256": "a7f2f338e5d4cf3faae6a9e2e13cb193ed71d019281d9b4cd4c8151e48d6ccb4",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"bin/fermion-run-linux-x64": {
	"bytes": 113296,
	"sha256": "39bc24eafc40f13594f63789b649e3698b326454c63ede118959e6e6197088d7",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"bin/fermion-run-linux-x64.sha256": {
	"bytes": 88,
	"sha256": "dcca4682d2f012df644caee9992ae8b48ab12ec0f1fb1a0ddc84aa865c530158",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"bin/fermion-run-macos-arm64": {
	"bytes": 583400,
	"sha256": "59b59d4df38baa08993a8429829c4889f1234183b1559b0078355f27351a1e66",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"bin/fermion-run-macos-arm64.sha256": {
	"bytes": 90,
	"sha256": "5664fda2394e3c3d1c8ca0a9869e4abe8367c307214a3070bd834457f74626da",
	"source": "2026-07-31 strict-parser rebuild"
	},
	"config.json": {
	"bytes": 501,
	"sha256": "589adacb286a009049219e4cb0b58537963cd7476e5ea59a66ef5c037802d05b",
	"source": "written for the trtc_v4 loader (hf_ternary registration)"
	},
	"generation_config.json": {
	"bytes": 181,
	"sha256": "2537909d4312b7155b60fe0aee96d342de1a00b7841ed3e7eba4c3835d16d8ae",
	"source": "transformers generation defaults for this SKU"
	},
	"gguf/Modelfile": {
	"bytes": 214,
	"sha256": "47c4c5ac0228e25197eb52e1ae0a964f95ac0d562d9effe9547f01c0cb58a9d1",
	"source": "Fermion Research release archive"
	},
	"gguf/README.md": {
	"bytes": 13234,
	"sha256": "029e36dd441be83bd70cad38f860760f8776fbb78a402ce7cb0361ae0048826b",
	"source": "banked lane artifact (updated 2026-07-31: Metal re-verify + CUDA rebuild + spec quickstart retarget)"
	},
	"gguf/SHA256SUMS": {
	"bytes": 745,
	"sha256": "8631ba8578ec9341a9367da853bf3d4979ef34e09ab3f22cac039bbd85ab03f7",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/bench_8b.json": {
	"bytes": 2663,
	"sha256": "1976449e3aa06f21e8337d0035ce8ee8371d6ff94d84adbafd149dbfa3f9d6fb",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/convert_8b.json": {
	"bytes": 953,
	"sha256": "09060b46412b45b856edd239c37b49208f7cebb153620dda1808296ace8308ea",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/gate_8b_fv5.json": {
	"bytes": 6989,
	"sha256": "393e6d48d74c27bca09ce82af855a1d5917347595477c4587774d0391b305e34",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/gateverdict_8b_fv5.json": {
	"bytes": 457,
	"sha256": "cd79e09be62309d4146da1530211529a9b98cffed0e240c7ef30c7466f8a64c1",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/ref_8b.json": {
	"bytes": 470,
	"sha256": "8916c4936da50ca500f521a19ce4d69668e8fd5603588113359efba5a059851a",
	"source": "Fermion Research release archive"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_bench.log": {
	"bytes": 1457,
	"sha256": "524fcf2783d38e7cf07b94c9d5ac0bdc73026e5b93aa447fe3fbd2d3927c0cbc",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_bench.txt": {
	"bytes": 572,
	"sha256": "05d67d1871a024a675f9921d50a0cf5693544a7dd240c83505e31717ac18baf4",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_bench_rerun.txt": {
	"bytes": 572,
	"sha256": "42f4627586d63fdae7c83bd788842daf94b0de01724241a1597712057516efdc",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_completion_metal.log": {
	"bytes": 2119,
	"sha256": "c564e46d1ff662dca8d7e8ac923efb3b4f41fb8409d56b6434ca724708c54e7e",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_completion_ngl0.log": {
	"bytes": 2128,
	"sha256": "528314eb5dedf1cb1cd94e44f7653ef6be42e2b72b1991e7ed285424532ba498",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_completion_ngl0.txt": {
	"bytes": 283,
	"sha256": "0bba48606aa39df95945023c750e75f2fd44772f6ca2226f93f4919f43503a26",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_testops_mulmat.txt": {
	"bytes": 501604,
	"sha256": "7d2dbd93d18137c3e6f89c062a98841e0de3af6dec260366c7753d85f9ac8c6e",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/after_testops_mulmat_cpu.txt": {
	"bytes": 175226,
	"sha256": "9b6b1890fbbd6a8abc396d8502c71f2bc5110dedf94f63e96b859272416f7454",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/before_bench.log": {
	"bytes": 1457,
	"sha256": "85f2a9c2e668a99c484c0cd97096d8d387a979a5d7e5320c91785af866476e2a",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/before_bench.txt": {
	"bytes": 572,
	"sha256": "7abeb886f9a4603fca71f77e010f2005b6fcf95aa3a3749e4a513586cc7b8a2d",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/before_bench_rerun.txt": {
	"bytes": 572,
	"sha256": "6fd1ecb171219a1f7acce27e7b3405d59814fa5f6a27f24d4ce0008e256bfc6f",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/before_completion_ngl0.log": {
	"bytes": 2120,
	"sha256": "4cc15a4f845e83cdf7cf2f0ef572669e2189ee5f4e623a123e245f32f454a929",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/before_completion_ngl0.txt": {
	"bytes": 283,
	"sha256": "0bba48606aa39df95945023c750e75f2fd44772f6ca2226f93f4919f43503a26",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/cpu_plain_k1.txt": {
	"bytes": 2486,
	"sha256": "fb01804820adc6a8a77042aa3ed004500afcbf26bd63b101ad0877566edb8c49",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/cpu_spec_k1.txt": {
	"bytes": 1848,
	"sha256": "bee5732a0090763f51c778a8a50f879faae97f28e004d3651954f7f92106a3b8",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"gguf/receipts/spec_cpu_m5_20260730/cpu_spec_stock.txt": {
	"bytes": 1848,
	"sha256": "d34be3aabcd80a30b52815ba3b94d6c9bfdd78c7713c8addc91e54904bae2c98",
	"source": "M5 CPU spec-kernel receipts (Apple M5, 2026-07-30): sgemm gates + k1 spec/plain runs"
	},
	"merges.txt": {
	"bytes": 1671853,
	"sha256": "8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5",
	"source": "Qwen/Qwen3-8B tokenizer file (Apache-2.0 upstream)"
	},
	"mlx/MANIFEST.json": {
	"bytes": 5951,
	"sha256": "d802e991e032c6a9eb841150b4a27c3f4b961bbe71563d7a3650bcb1492d54e6",
	"source": "banked lane artifact (regenerated 2026-07-31: v10 fermion_mlx pack pins)"
	},
	"mlx/README.md": {
	"bytes": 5998,
	"sha256": "d4327ce1c08545b05a4a541c38d566a3958e7634869f85d57ded3b726869c84b",
	"source": "Fermion Research release archive"
	},
	"mlx/fermion_mlx/MANIFEST.json": {
	"bytes": 1077,
	"sha256": "6f2f11bf17214c7f450f231b377a9b47c446904e86ea670cdce9b328cf472cb8",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/__init__.py": {
	"bytes": 6626,
	"sha256": "999f003da4e49311413876df3b4011eae515fbf15684a96a4794af8f946d5185",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/__main__.py": {
	"bytes": 13847,
	"sha256": "8ba4104366a29737b5fecc042ecfac88eb0106122464af29c5822b2ee81ba7d9",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/arch3.py": {
	"bytes": 8206,
	"sha256": "84a15da8566acdfbf6acef60e09099113db150c0f53deefb04491365ea7efef5",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/container.py": {
	"bytes": 6059,
	"sha256": "c5b352184ba9ddf9a3e424731226cb7dcb9901d3a1516c67d4557953b0101f2f",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/drafter.py": {
	"bytes": 6892,
	"sha256": "a0c5ace9e4dc94926f46e6495407e9463676d25345884a6eef626dcb92a9de14",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/flash_sdpa.py": {
	"bytes": 11367,
	"sha256": "6ce98b11049ee902db14259b3db320f7df32b9be4a1fb13d86a2a76577a30cb1",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/gates.py": {
	"bytes": 7411,
	"sha256": "c032e35944e4ec3c0b991e246114697a9538fa9bfcfff183d39b066d3d180349",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/heads.py": {
	"bytes": 9239,
	"sha256": "a9353d3e9bd93130a1533379986b11c7baefa0e11af555c36a94d9f9d0e97226",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/mbatch.py": {
	"bytes": 20068,
	"sha256": "5736d2cf0ef3d7215b92beebc2c4dc3fbdc09f147effd5f7e3b8b696cdfb71a1",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/mma.py": {
	"bytes": 14725,
	"sha256": "82905730ea79240d9647e63a8131c076781e72979efe2a05999c911206db8c96",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/model.py": {
	"bytes": 26179,
	"sha256": "7d48f63b7043c8e8508a937f0d2e2e4a17a82c02800a082671a7dbae163d0d72",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/spec.py": {
	"bytes": 15875,
	"sha256": "cb92be27e752580f3e9852191d50dc6b421cd5e33d18b7f53b528d4a832c7895",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/fermion_mlx/trit_kernels.py": {
	"bytes": 44235,
	"sha256": "ba1c32e18585d8a1e6202c739ba6e4cbcdaefcc23f0dd7d665ffbfcfb7bbd1d4",
	"source": "v10 MLX pack rebuild 2026-07-31"
	},
	"mlx/merges.txt": {
	"bytes": 1671853,
	"sha256": "8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5",
	"source": "Fermion Research release archive"
	},
	"mlx/receipts/pathgate_8b_cons3b25.json": {
	"bytes": 634,
	"sha256": "ff7f68d926d381a6c1a00f07cbb33ad5b3ac118d7b9dd997de0857cc1af40f71",
	"source": "Fermion Research release archive"
	},
	"mlx/receipts/pathgate_8b_product.json": {
	"bytes": 633,
	"sha256": "c13e35c461c6e3a23fb2976d0f26f0b8dce114b12bf976a05cb19f11cc0746ea",
	"source": "Fermion Research release archive"
	},
	"mlx/receipts/refgate_8b_cons3b25.json": {
	"bytes": 1938,
	"sha256": "bc35c97e941f3f95f8fec3466c60c27f5fe96b681cfe165c53f323c7fad8130a",
	"source": "Fermion Research release archive"
	},
	"mlx/receipts/specbench_8b_0p6b_m5.json": {
	"bytes": 826,
	"sha256": "c36ef09659d8c7690d53c23d065943103a95417890f8defbd01e629d7b5feeca",
	"source": "Fermion Research release archive"
	},
	"mlx/receipts/specgate_8b_0p6b.json": {
	"bytes": 705,
	"sha256": "64bcb90fe965835bc78887d873d3cfd0677b76c6d72cda04a84e295efee7c10c",
	"source": "Fermion Research release archive"
	},
	"mlx/requirements.txt": {
	"bytes": 191,
	"sha256": "1cf4d1fa581aaa5deae0568f88388c51166bbd69e7a89080b5a53488a7ef39b6",
	"source": "Fermion Research release archive"
	},
	"mlx/tokenizer.json": {
	"bytes": 11422654,
	"sha256": "aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4",
	"source": "Fermion Research release archive"
	},
	"mlx/tokenizer_config.json": {
	"bytes": 9980,
	"sha256": "9f1e48d00384f25decc7f30dc7ec3a76fb3a98b37f5297b8494a134cef786102",
	"source": "Fermion Research release archive"
	},
	"mlx/vocab.json": {
	"bytes": 2776833,
	"sha256": "ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910",
	"source": "Fermion Research release archive"
	},
	"receipts/draft/acceptance_20260730.log": {
	"bytes": 3766,
	"sha256": "03887954f5c17bc25f0d15a883a84b87baf8f3f2aed10eeb730401c9d8cbcb5d",
	"source": "draft per-class acceptance probe log (H100 2026-07-30) - scrubbed public copy"
	},
	"receipts/draft/cert_20260730.log": {
	"bytes": 12408,
	"sha256": "02410533f096b561ad61f374d35b295f60f96567d066d70d854a4718dca0cf6c",
	"source": "draft certification log (27,648 spec tokens, 0 mismatches; H100 2026-07-30) - scrubbed public copy"
	},
	"receipts/dspec_cert_cons3_b25.json": {
	"bytes": 4696,
	"sha256": "2653fbad0cca5277ee4fa3e74360f8900f093b8e38b5cb3e67047e7dc87c88d9",
	"source": "Modal H100 speculative-decoding exactness certificate for the SHIPPING container"
	},
	"receipts/mlx_spec_door_20260730.json": {
	"bytes": 980,
	"sha256": "a080c07c321e3b46dd90d35d1363906977ce63e1b965bbdf87d2961d461373b6",
	"source": "MLX spec-door bench receipt (Apple M5), 2026-07-30"
	},
	"receipts/rebuild_gate/lane_a.json": {
	"bytes": 12323,
	"sha256": "d2ba025facc445fdc64c91dbc2091ed91c181856c881a52a628c511b3d0d02a4",
	"source": "rebuild-gate receipt, lane A (NVIDIA L4), 2026-07-30 - scrubbed public copy"
	},
	"receipts/rebuild_gate/lane_b.json": {
	"bytes": 2433,
	"sha256": "cd0580231a6a99d8ed225a703d7452688389936e871163ff474a173d3b096d3d",
	"source": "rebuild-gate receipt, lane B (Linux x86-64 CPU), 2026-07-30 - scrubbed public copy"
	},
	"tokenizer.json": {
	"bytes": 11422654,
	"sha256": "aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4",
	"source": "Qwen/Qwen3-8B tokenizer file (Apache-2.0 upstream)"
	},
	"tokenizer_config.json": {
	"bytes": 9979,
	"sha256": "02e7044c846290e910d9c9deca9812315d64be56a6aeee3cfbedc99d177644c4",
	"source": "Qwen/Qwen3-8B tokenizer file (Apache-2.0 upstream)"
	},
	"vocab.json": {
	"bytes": 2776833,
	"sha256": "ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910",
	"source": "Qwen/Qwen3-8B tokenizer file (Apache-2.0 upstream)"
	},
	"receipts/rebuild_gate/macos_gate_20260731.json": {
	"bytes": 1282,
	"sha256": "f40dafaed148429dd2379362dc8ac4f474f225cecfe3c932de7ebda6408be5a0",
	"source": "M5 + Modal cross-platform gate 2026-07-31 (identity/determinism/spec-exactness)"
	},
	"mlx/receipts/noflash_fp16_race_20260731.json": {
	"bytes": 1287,
	"sha256": "efd6ad55ee8d994ba509347542ad49c409d78372deb949dcd75f8fe8bac94a67",
	"source": "no-flash fallback race autopsy + ship constraint (2026-07-31)"
	},
	"receipts/rebuild_gate/linux_regate_20260731.json": {
	"bytes": 5396,
	"sha256": "3f324683603c64fb5dd3757d28fe5a85df4518704d1d93fab57579b6a39e7828",
	"source": "Modal linux regate 7/7 (2026-07-31)"
	}
	}
	}