Instructions to use singularity-layer/Singularity-xMESH-0.8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use singularity-layer/Singularity-xMESH-0.8B with Transformers:

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

pipe = pipeline("image-text-to-text", model="singularity-layer/Singularity-xMESH-0.8B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("singularity-layer/Singularity-xMESH-0.8B")
model = AutoModelForMultimodalLM.from_pretrained("singularity-layer/Singularity-xMESH-0.8B", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use singularity-layer/Singularity-xMESH-0.8B 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 singularity-layer/Singularity-xMESH-0.8B:BF16
# Run inference directly in the terminal:
llama cli -hf singularity-layer/Singularity-xMESH-0.8B:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf singularity-layer/Singularity-xMESH-0.8B:BF16
# Run inference directly in the terminal:
llama cli -hf singularity-layer/Singularity-xMESH-0.8B:BF16

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 singularity-layer/Singularity-xMESH-0.8B:BF16
# Run inference directly in the terminal:
./llama-cli -hf singularity-layer/Singularity-xMESH-0.8B:BF16

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 singularity-layer/Singularity-xMESH-0.8B:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf singularity-layer/Singularity-xMESH-0.8B:BF16

Use Docker

docker model run hf.co/singularity-layer/Singularity-xMESH-0.8B:BF16

LM Studio
Jan

vLLM

How to use singularity-layer/Singularity-xMESH-0.8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "singularity-layer/Singularity-xMESH-0.8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "singularity-layer/Singularity-xMESH-0.8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/singularity-layer/Singularity-xMESH-0.8B:BF16

SGLang

How to use singularity-layer/Singularity-xMESH-0.8B 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 "singularity-layer/Singularity-xMESH-0.8B" \
    --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": "singularity-layer/Singularity-xMESH-0.8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "singularity-layer/Singularity-xMESH-0.8B" \
        --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": "singularity-layer/Singularity-xMESH-0.8B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Ollama
How to use singularity-layer/Singularity-xMESH-0.8B with Ollama:
```
ollama run hf.co/singularity-layer/Singularity-xMESH-0.8B:BF16
```
Unsloth Desktop

How to use singularity-layer/Singularity-xMESH-0.8B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf singularity-layer/Singularity-xMESH-0.8B:BF16

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "singularity-layer/Singularity-xMESH-0.8B:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use singularity-layer/Singularity-xMESH-0.8B with Docker Model Runner:
```
docker model run hf.co/singularity-layer/Singularity-xMESH-0.8B:BF16
```

Lemonade

How to use singularity-layer/Singularity-xMESH-0.8B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull singularity-layer/Singularity-xMESH-0.8B:BF16

Run and chat with the model

lemonade run user.Singularity-xMESH-0.8B-BF16

List all available models

lemonade list

Hermes Agent

How to use singularity-layer/Singularity-xMESH-0.8B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf singularity-layer/Singularity-xMESH-0.8B:BF16

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 singularity-layer/Singularity-xMESH-0.8B:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use singularity-layer/Singularity-xMESH-0.8B with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf singularity-layer/Singularity-xMESH-0.8B:BF16

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 "singularity-layer/Singularity-xMESH-0.8B:BF16" \
  --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"

Singularity-xMESH-0.8B / README.md

ivaavi

Link the release article and Singularity Layer website

4026d47 verified 1 day ago

preview code

Raw History Blame Contribute Delete

8.42 kB

---
license: apache-2.0
base_model:
  - Qwen/Qwen3.5-0.8B
datasets:
  - MaziyarPanahi/AgentToolDecisions-180K
library_name: transformers
tags:
  - singularity-layer
  - xmesh
  - decision-model
  - x402
  - tool-selection
  - safetensors
  - gguf
  - llama.cpp
  - pytorch
  - cpu
  - apple-silicon
  - windows
  - linux
  - mps
  - cuda
language:
  - en
---
<picture>
  <source media="(prefers-color-scheme: dark)" srcset="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-light.png">
  <source media="(prefers-color-scheme: light)" srcset="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-dark.png">
  <img alt="Singularity xMESH 0.8B" src="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-dark.png" width="100%">
</picture>

# Singularity xMESH 0.8B

**x402 Multi-Endpoint Selection Head**

**Trained and evaluated with Singularity Training Lab.**

**[Release article](https://www.x402layer.cc/blog/singularity-xmesh-0-8b) · [Singularity Layer](https://www.x402layer.cc)**

Singularity xMESH 0.8B is a compact decision model for selecting services and composing constrained multi-step x402 flows. This repository contains the complete standalone model, its calibrated joint decision head, and the deterministic xMESH policy. No separate base-model download is required.

## Intended behavior

Input: an intent, hard constraints, and a small candidate set retrieved from a current x402 catalog. The validated runtime contract is 2,048 tokens.

Output: candidate selection or abstention, calibrated confidence, approval requirement, network compatibility, and cost level. Deterministic policy validates stage gates and flow feasibility before any paid request. Service choices remain model decisions.

## Training data

The general decision warm-up used licensed rows from [`MaziyarPanahi/AgentToolDecisions-180K`](https://huggingface.co/datasets/MaziyarPanahi/AgentToolDecisions-180K). That dataset retains mixed CC BY 4.0, Apache-2.0, and MIT source terms per row; see `THIRD_PARTY_NOTICES.md`. Singularity-curated x402 rows supplied the routing specialization. Frozen evaluation cases and resources were excluded from training.

## Hardware compatibility

| Hardware | Status |
| --- | --- |
| Apple silicon MPS | **Verified** on an M4 with 16 GB unified memory |
| CPU | **Verified**; portable but slower |
| llama.cpp Metal | **Verified** with BF16 full offload on an Apple M4: 959/960 frozen-suite decisions match CPU (one near-tie). Q8 also passes 960/960 |
| llama.cpp Windows/Linux x64 CPU | **Verified**; native canaries passed 9/9 expected decisions on both platforms |
| llama.cpp Vulkan on NVIDIA | **Verified** on an RTX A4000: 960/960 frozen-suite decisions match CPU, 0.31 s median per request |
| NVIDIA CUDA | PyTorch runtime support; the llama.cpp CUDA build is not yet verified |
| AMD and Intel GPUs | Not verified; Singularity Node runs xMESH on CPU there |

Singularity Node v1.10.4 and later choose the path automatically: full GPU offload on Apple silicon and on NVIDIA with the Vulkan build, CPU everywhere else. Full results: `ACCELERATOR_PARITY.json`.

The checkpoint contains 873M BF16 parameters. An 8 GB system is the practical minimum; 16 GB is recommended for comfortable local loading and inference.

## Portable GGUF runtime

`Singularity-xMESH-0.8B-BF16.gguf` is the universal Singularity Node artifact. It is 1,571,529,600 bytes with SHA-256 `9f4c08f19c4d203059bc74ed2754e800ec472d7276d8ddb2321480e86d392fff` and requires llama.cpp `b11719` or newer for native Clef `/v1/systemone` support.

`Singularity-xMESH-0.8B-Q8_0.gguf` is an optional 841,069,440-byte accelerated artifact with SHA-256 `41b36270b67f6c595e98cf4f29d6a0d13f0c19daa37b75741cd8296e2729fd75`. It preserved all 960 decisions across the frozen 80-case Metal export-parity run, but llama.cpp `b11719` crashes while initializing this quantized Clef architecture on CPU. Singularity Node therefore uses BF16 for its cross-platform contract. Q4_K_M was rejected because it changed 16 decisions and preserved only 65 of 80 complete flows.

## Frozen x402 regression evaluation

The standalone export was fixed before one evaluation on the existing frozen 80-case x402 suite. No weights, prompts, or policy rules were changed after seeing these results.

| Evaluation, 80 cases | xMESH standalone | Jev | Laya |
| --- | ---: | ---: | ---: |
| Complete-flow exact match | **80.0%** | 11.25% | 0.0% |
| Stage-gate accuracy | **96.25%** | 86.5% | 78.5% |
| Active-service accuracy | 97.5% | **100.0%** | 85.63% |
| Global-decision accuracy | **96.25%** | 37.5% | 35.63% |
| Executed cost regret | $0.00175 | **$0.00000** | $0.00358 |
| Post-policy violations | **0** | — | — |

Jev is a hosted API and Laya used a different local runtime, so latency is not a hardware-equivalent comparison. The standalone CPU validation measured 6.45 seconds median per full multi-field flow on an Apple M4. Optimized serving runtimes can be substantially faster.

## General typed-decision evaluation

This is a boundary test, not the model's specialization. xMESH was evaluated zero-shot on `LocalLLaMA/typed-decisions` revision `7189f46b02248bd2569df7325c9df8e3b24e90a7`, complete `all/test` split: 400 cases and 2,000 decisions.

| Model | Condition | Accuracy | Soft accuracy | Brier ↓ | ECE ↓ | Score MAE ↓ |
| --- | --- | ---: | ---: | ---: | ---: | ---: |
| xMESH 0.8B | zero-shot Q8 | **0.478** | 0.415 | 0.409 | 0.293 | 0.642 |
| Laya | base checkpoint, published | 0.362 | 0.331 | **0.315** | **0.174** | 0.694 |
| Laya typed-decisions | trained on this benchmark's train split, published | **0.766** | 0.471 | **0.061** | 0.213 | **0.242** |
| Jev 1.13 | published | 0.727 | **0.580** | 0.148 | **0.144** | 0.391 |

xMESH beats base Laya on this suite, but it does not beat Jev or the benchmark-specific Laya fine-tune. Accuracy by xMESH workflow was 0.266 agent-trace observability, 0.578 customer service, 0.492 invoice processing, and 0.574 security incidents.

A seeded option-order permutation changed **0 of 600** choice answers and produced zero probability drift because the Clef runtime canonicalizes option identifiers before inference.

On a short fixed state with identical three-option questions, Q8 llama.cpp on an Apple M4 measured 168 ms p50 for one question, 448 ms for five, and 847 ms for ten. Twenty and fifty questions exceeded this release's validated 2,048-token context; they are not advertised as supported.

Laya and Jev comparison rows above are figures published by the Laya benchmark page. They were not rerun in the xMESH environment.

## Run on Apple silicon

Validated with Python 3.12 on an Apple M4 with 16 GB unified memory.

```bash
python3.12 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python examples/inference.py --request examples/request.json --device mps
```

Use `--device cpu` when MPS is unavailable. The request must contain `state` and `questions`; `examples/request.json` shows the typed contract.

## Run with llama.cpp

```bash
llama-server \
  -m Singularity-xMESH-0.8B-BF16.gguf \
  --host 127.0.0.1 --port 8081 \
  -c 2048 -b 2048 -ub 2048 --parallel 1 -ngl 0
curl http://127.0.0.1:8081/v1/systemone \
  -H 'Content-Type: application/json' \
  --data-binary @examples/request.json
```

`-ngl 0` is the portable CPU setting. On Apple silicon (Metal) or an NVIDIA GPU with the Vulkan build, `-ngl 99` is verified to give the same decisions; see `ACCELERATOR_PARITY.json`.

## Safety boundary

xMESH is an advisory router. It must not receive wallet secrets, construct signed payments, bypass approval, invent endpoints, or execute arbitrary URLs. Runtime code must restrict choices to retrieved candidate IDs and enforce budgets, networks, schemas, and approval rules independently.

## Limitations

- Specialized for x402 service routing; it is not a general chat model.
- It does not retrieve services, verify providers, quote payments, or execute requests.
- Service-selection accuracy is strong but not perfect.
- Confidence is advisory and does not replace deterministic policy checks.
- The validated context contract is 2,048 tokens.

## License

The model and Singularity runtime files are Apache-2.0. Qwen and Cloudflare Clef attribution is included in `THIRD_PARTY_NOTICES.md`.

	---
	license: apache-2.0
	base_model:
	- Qwen/Qwen3.5-0.8B
	datasets:
	- MaziyarPanahi/AgentToolDecisions-180K
	library_name: transformers
	tags:
	- singularity-layer
	- xmesh
	- decision-model
	- x402
	- tool-selection
	- safetensors
	- gguf
	- llama.cpp
	- pytorch
	- cpu
	- apple-silicon
	- windows
	- linux
	- mps
	- cuda
	language:
	- en
	---

	<picture>
	<source media="(prefers-color-scheme: dark)" srcset="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-light.png">
	<source media="(prefers-color-scheme: light)" srcset="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-dark.png">
	<img alt="Singularity xMESH 0.8B" src="https://huggingface.co/singularity-layer/Singularity-xMESH-0.8B/resolve/main/assets/xmesh-wordmark-dark.png" width="100%">
	</picture>

	# Singularity xMESH 0.8B

	x402 Multi-Endpoint Selection Head

	Trained and evaluated with Singularity Training Lab.

	[Release article](https://www.x402layer.cc/blog/singularity-xmesh-0-8b) · [Singularity Layer](https://www.x402layer.cc)

	Singularity xMESH 0.8B is a compact decision model for selecting services and composing constrained multi-step x402 flows. This repository contains the complete standalone model, its calibrated joint decision head, and the deterministic xMESH policy. No separate base-model download is required.

	## Intended behavior

	Input: an intent, hard constraints, and a small candidate set retrieved from a current x402 catalog. The validated runtime contract is 2,048 tokens.

	Output: candidate selection or abstention, calibrated confidence, approval requirement, network compatibility, and cost level. Deterministic policy validates stage gates and flow feasibility before any paid request. Service choices remain model decisions.

	## Training data

	The general decision warm-up used licensed rows from [`MaziyarPanahi/AgentToolDecisions-180K`](https://huggingface.co/datasets/MaziyarPanahi/AgentToolDecisions-180K). That dataset retains mixed CC BY 4.0, Apache-2.0, and MIT source terms per row; see `THIRD_PARTY_NOTICES.md`. Singularity-curated x402 rows supplied the routing specialization. Frozen evaluation cases and resources were excluded from training.

	## Hardware compatibility

	\| Hardware \| Status \|
	\| --- \| --- \|
	\| Apple silicon MPS \| Verified on an M4 with 16 GB unified memory \|
	\| CPU \| Verified; portable but slower \|
	\| llama.cpp Metal \| Verified with BF16 full offload on an Apple M4: 959/960 frozen-suite decisions match CPU (one near-tie). Q8 also passes 960/960 \|
	\| llama.cpp Windows/Linux x64 CPU \| Verified; native canaries passed 9/9 expected decisions on both platforms \|
	\| llama.cpp Vulkan on NVIDIA \| Verified on an RTX A4000: 960/960 frozen-suite decisions match CPU, 0.31 s median per request \|
	\| NVIDIA CUDA \| PyTorch runtime support; the llama.cpp CUDA build is not yet verified \|
	\| AMD and Intel GPUs \| Not verified; Singularity Node runs xMESH on CPU there \|

	Singularity Node v1.10.4 and later choose the path automatically: full GPU offload on Apple silicon and on NVIDIA with the Vulkan build, CPU everywhere else. Full results: `ACCELERATOR_PARITY.json`.

	The checkpoint contains 873M BF16 parameters. An 8 GB system is the practical minimum; 16 GB is recommended for comfortable local loading and inference.

	## Portable GGUF runtime

	`Singularity-xMESH-0.8B-BF16.gguf` is the universal Singularity Node artifact. It is 1,571,529,600 bytes with SHA-256 `9f4c08f19c4d203059bc74ed2754e800ec472d7276d8ddb2321480e86d392fff` and requires llama.cpp `b11719` or newer for native Clef `/v1/systemone` support.

	`Singularity-xMESH-0.8B-Q8_0.gguf` is an optional 841,069,440-byte accelerated artifact with SHA-256 `41b36270b67f6c595e98cf4f29d6a0d13f0c19daa37b75741cd8296e2729fd75`. It preserved all 960 decisions across the frozen 80-case Metal export-parity run, but llama.cpp `b11719` crashes while initializing this quantized Clef architecture on CPU. Singularity Node therefore uses BF16 for its cross-platform contract. Q4_K_M was rejected because it changed 16 decisions and preserved only 65 of 80 complete flows.

	## Frozen x402 regression evaluation

	The standalone export was fixed before one evaluation on the existing frozen 80-case x402 suite. No weights, prompts, or policy rules were changed after seeing these results.

	\| Evaluation, 80 cases \| xMESH standalone \| Jev \| Laya \|
	\| --- \| ---: \| ---: \| ---: \|
	\| Complete-flow exact match \| 80.0% \| 11.25% \| 0.0% \|
	\| Stage-gate accuracy \| 96.25% \| 86.5% \| 78.5% \|
	\| Active-service accuracy \| 97.5% \| 100.0% \| 85.63% \|
	\| Global-decision accuracy \| 96.25% \| 37.5% \| 35.63% \|
	\| Executed cost regret \| $0.00175 \| $0.00000 \| $0.00358 \|
	\| Post-policy violations \| 0 \| — \| — \|

	Jev is a hosted API and Laya used a different local runtime, so latency is not a hardware-equivalent comparison. The standalone CPU validation measured 6.45 seconds median per full multi-field flow on an Apple M4. Optimized serving runtimes can be substantially faster.

	## General typed-decision evaluation

	This is a boundary test, not the model's specialization. xMESH was evaluated zero-shot on `LocalLLaMA/typed-decisions` revision `7189f46b02248bd2569df7325c9df8e3b24e90a7`, complete `all/test` split: 400 cases and 2,000 decisions.

	\| Model \| Condition \| Accuracy \| Soft accuracy \| Brier ↓ \| ECE ↓ \| Score MAE ↓ \|
	\| --- \| --- \| ---: \| ---: \| ---: \| ---: \| ---: \|
	\| xMESH 0.8B \| zero-shot Q8 \| 0.478 \| 0.415 \| 0.409 \| 0.293 \| 0.642 \|
	\| Laya \| base checkpoint, published \| 0.362 \| 0.331 \| 0.315 \| 0.174 \| 0.694 \|
	\| Laya typed-decisions \| trained on this benchmark's train split, published \| 0.766 \| 0.471 \| 0.061 \| 0.213 \| 0.242 \|
	\| Jev 1.13 \| published \| 0.727 \| 0.580 \| 0.148 \| 0.144 \| 0.391 \|

	xMESH beats base Laya on this suite, but it does not beat Jev or the benchmark-specific Laya fine-tune. Accuracy by xMESH workflow was 0.266 agent-trace observability, 0.578 customer service, 0.492 invoice processing, and 0.574 security incidents.

	A seeded option-order permutation changed 0 of 600 choice answers and produced zero probability drift because the Clef runtime canonicalizes option identifiers before inference.

	On a short fixed state with identical three-option questions, Q8 llama.cpp on an Apple M4 measured 168 ms p50 for one question, 448 ms for five, and 847 ms for ten. Twenty and fifty questions exceeded this release's validated 2,048-token context; they are not advertised as supported.

	Laya and Jev comparison rows above are figures published by the Laya benchmark page. They were not rerun in the xMESH environment.

	## Run on Apple silicon

	Validated with Python 3.12 on an Apple M4 with 16 GB unified memory.

	```bash
	python3.12 -m venv .venv
	source .venv/bin/activate
	pip install -r requirements.txt
	python examples/inference.py --request examples/request.json --device mps
	```

	Use `--device cpu` when MPS is unavailable. The request must contain `state` and `questions`; `examples/request.json` shows the typed contract.

	## Run with llama.cpp

	```bash
	llama-server \
	-m Singularity-xMESH-0.8B-BF16.gguf \
	--host 127.0.0.1 --port 8081 \
	-c 2048 -b 2048 -ub 2048 --parallel 1 -ngl 0

	curl http://127.0.0.1:8081/v1/systemone \
	-H 'Content-Type: application/json' \
	--data-binary @examples/request.json
	```

	`-ngl 0` is the portable CPU setting. On Apple silicon (Metal) or an NVIDIA GPU with the Vulkan build, `-ngl 99` is verified to give the same decisions; see `ACCELERATOR_PARITY.json`.

	## Safety boundary

	xMESH is an advisory router. It must not receive wallet secrets, construct signed payments, bypass approval, invent endpoints, or execute arbitrary URLs. Runtime code must restrict choices to retrieved candidate IDs and enforce budgets, networks, schemas, and approval rules independently.

	## Limitations

	- Specialized for x402 service routing; it is not a general chat model.
	- It does not retrieve services, verify providers, quote payments, or execute requests.
	- Service-selection accuracy is strong but not perfect.
	- Confidence is advisory and does not replace deterministic policy checks.
	- The validated context contract is 2,048 tokens.

	## License

	The model and Singularity runtime files are Apache-2.0. Qwen and Cloudflare Clef attribution is included in `THIRD_PARTY_NOTICES.md`.