Instructions to use wesleysimplicio/Simplicio-27B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wesleysimplicio/Simplicio-27B with Transformers:

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

pipe = pipeline("text-generation", model="wesleysimplicio/Simplicio-27B")
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("wesleysimplicio/Simplicio-27B")
model = AutoModelForMultimodalLM.from_pretrained("wesleysimplicio/Simplicio-27B", 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 wesleysimplicio/Simplicio-27B 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 wesleysimplicio/Simplicio-27B:BF16
# Run inference directly in the terminal:
llama cli -hf wesleysimplicio/Simplicio-27B:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf wesleysimplicio/Simplicio-27B:BF16
# Run inference directly in the terminal:
llama cli -hf wesleysimplicio/Simplicio-27B:BF16

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf wesleysimplicio/Simplicio-27B:BF16
# Run inference directly in the terminal:
./build/bin/llama cli -hf wesleysimplicio/Simplicio-27B:BF16

Use Docker

docker model run hf.co/wesleysimplicio/Simplicio-27B:BF16

Llama new
LM Studio
Jan

vLLM

How to use wesleysimplicio/Simplicio-27B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "wesleysimplicio/Simplicio-27B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wesleysimplicio/Simplicio-27B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/wesleysimplicio/Simplicio-27B:BF16

SGLang

How to use wesleysimplicio/Simplicio-27B 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 "wesleysimplicio/Simplicio-27B" \
    --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": "wesleysimplicio/Simplicio-27B",
		"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 "wesleysimplicio/Simplicio-27B" \
        --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": "wesleysimplicio/Simplicio-27B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use wesleysimplicio/Simplicio-27B with Ollama:
```
ollama run hf.co/wesleysimplicio/Simplicio-27B:BF16
```
Unsloth Desktop

How to use wesleysimplicio/Simplicio-27B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf wesleysimplicio/Simplicio-27B: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": "wesleysimplicio/Simplicio-27B:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use wesleysimplicio/Simplicio-27B with Docker Model Runner:
```
docker model run hf.co/wesleysimplicio/Simplicio-27B:BF16
```

Lemonade

How to use wesleysimplicio/Simplicio-27B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull wesleysimplicio/Simplicio-27B:BF16

Run and chat with the model

lemonade run user.Simplicio-27B-BF16

List all available models

lemonade list

Hermes Agent

How to use wesleysimplicio/Simplicio-27B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf wesleysimplicio/Simplicio-27B: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 wesleysimplicio/Simplicio-27B:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use wesleysimplicio/Simplicio-27B with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf wesleysimplicio/Simplicio-27B: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 "wesleysimplicio/Simplicio-27B: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"

wesleysimplicio commited on 27 days ago

Commit

c0fef9d

verified ·

1 Parent(s): b7c2ddb

docs: benchmark Simplicio 27B against 2026 generation frontier models (Claude 3.7, o3, o3-mini, DeepSeek-R1, Gemini 2.0 Pro)

Browse files

Files changed (1) hide show

README.md +22 -22

README.md CHANGED Viewed

@@ -68,41 +68,41 @@ Unlike standard conversational models that employ unbounded, verbose Chain-of-Th
 ---
-## 🏆 Top 15 Coding & Agentic Software Engineering LLMs Leaderboard (State-of-the-Art)
-The leaderboard below synthesizes the state-of-the-art across the **15 premier contemporary models for Autonomous Software Engineering and Coding Agents** (including **Claude 3.7 Sonnet**, **OpenAI o3-mini**, **DeepSeek-R1**, **Gemini 2.0 Flash**, and **Llama 3.3 70B**), benchmarked on **Surgical Diff Accuracy (Aider Benchmark)**, **SWE-bench Verified**, and **Reasoning Token Consumption**.
 <p align="center">
-  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/tier_list_coding.svg" alt="Software Engineering &amp; Coding Tier List (Latest Models)" width="100%">
 </p>
 <p align="center">
-  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/leaderboard_top15.svg" alt="Top 15 Coding LLMs Leaderboard (Latest Models)" width="100%">
 </p>
 <p align="center">
-  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/token_economy_cost.svg" alt="Reasoning Efficiency: Useful Code vs. Expensive Chain-of-Thought" width="100%">
 </p>
-### 📊 Comparative Scorecard: Top 15 Contemporary Models
-| Rank | Model | Creator / Org | Architecture / Size | Type | Surgical Diff (Aider) | SWE-bench Verified | Tokens / Task (Lower is better) | Core Superpower / Highlight |
 | :---: | :--- | :---: | :---: | :---: | :---: | :---: | :---: | :--- |
-| 🥇 **#1** | **Claude 3.7 Sonnet (Thinking)** | Anthropic | Frontier Hybrid | 🔒 Closed | 88.0% | **70.3%** | 1,400 t | State-of-the-art hybrid reasoning &amp; SWE-bench |
-| 🥈 **#2** | **⚡ Simplicio 27B** | simpletibr | **27B Hybrid (DeltaNet)** | 🟢 **Open** | **96.5%** | **46.5%** | **480 t** *(⚡ -43% economy)* | **#1 in Atomic Surgical Diff &amp; 50-Point Loop Protocol** |
-| 🥉 **#3** | **OpenAI o3-mini (High Effort)** | OpenAI | Frontier Reasoning | 🔒 Closed | 82.0% | 53.0% | 1,650 t | Deep algorithmic reasoning and strict logic |
-| **#4** | **Claude 3.5 Sonnet (v2)** | Anthropic | Frontier MoE | 🔒 Closed | 84.0% | 49.2% | 680 t | Repository-level issue resolution and UI vision |
-| **#5** | **DeepSeek-R1 (Full 671B)** | DeepSeek | 671B MoE CoT | 🟢 Open | 76.5% | 49.2% | 1,850 t | Unsupervised open reasoning trained via RL |
-| **#6** | **DeepSeek-V3** | DeepSeek | 671B MoE (37B active) | 🟢 Open | 75.0% | 43.4% | 790 t | High knowledge density and exceptional cost profile |
-| **#7** | **OpenAI o1** | OpenAI | Frontier Reasoning | 🔒 Closed | 78.0% | 48.9% | 2,100 t | Deep Chain-of-Thought reasoning pioneer |
-| **#8** | **Gemini 2.0 Flash (Thinking)** | Google | Frontier Multimodal | 🔒 Closed | 74.5% | 42.0% | 920 t | Ultra-low latency and massive multimodal context |
-| **#9** | **GPT-4o (Latest)** | OpenAI | Frontier Dense | 🔒 Closed | 73.5% | 38.8% | 720 t | Generalist stability, tool routing, and API maturity |
-| **#10** | **Qwen 2.5 Coder 32B** | Alibaba | 32B Dense | 🟢 Open | 73.0% | 39.8% | 740 t | Best dense pre-trained base model for code completion |
-| **#11** | **Llama 3.3 70B Instruct** | Meta | 70B Dense | 🟢 Open | 67.5% | 37.5% | 790 t | Meta's flagship open-weights enterprise baseline |
-| **#12** | **Qwen3.8-27B (Thinking)** | Alibaba | 27B Hybrid DeltaNet | 🟢 Open | 69.2% | 40.8% | 850 t | Native DeltaNet hybrid reasoning with verbose CoT |
-| **#13** | **Codestral 25.01** | Mistral | 24B Dense | 🟢 Open | 68.0% | 33.0% | 660 t | Mistral's latest IDE and fill-in-the-middle model |
-| **#14** | **Qwen3.8-27B (Fast)** | Alibaba | 27B Hybrid DeltaNet | 🟢 Open | 58.4% | 33.5% | 590 t | Ultra-fast generation prone to diff formatting errors |
-| **#15** | **Yi-Coder 9B** | 01.AI | 9B Dense | 🟢 Open | 52.0% | 23.5% | 650 t | Lightweight model for resource-constrained environments |
 ---

 ---
+## 🏆 2026 Top 15 Coding & Agentic Software Engineering LLMs Leaderboard
+The benchmark below evaluates the **15 state-of-the-art models of the 2026 generation for Autonomous Software Engineering and Coding Agents** (including **Claude 3.7 Sonnet**, **OpenAI o3**, **OpenAI o3-mini**, **DeepSeek-R1**, **Gemini 2.0 Pro**, and **Llama 3.3 70B**), benchmarked on **Surgical Diff Accuracy (Aider Benchmark)**, **SWE-bench Verified**, and **Reasoning Token Consumption**.
 <p align="center">
+  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/tier_list_coding.svg" alt="2026 Software Engineering &amp; Coding Tier List" width="100%">
 </p>
 <p align="center">
+  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/leaderboard_top15.svg" alt="2026 Top 15 Coding LLMs Leaderboard" width="100%">
 </p>
 <p align="center">
+  <img src="https://raw.githubusercontent.com/simpletibr/simplicio-27b/main/assets/token_economy_cost.svg" alt="2026 Reasoning Efficiency: Useful Code vs. Expensive Chain-of-Thought" width="100%">
 </p>
+### 📊 Comparative Scorecard: 2026 Model Generation
+| Rank | 2026 Model | Creator / Org | Architecture / Size | Type | Surgical Diff (Aider) | SWE-bench Verified | Tokens / Task (Lower is better) | Core Superpower / Highlight |
 | :---: | :--- | :---: | :---: | :---: | :---: | :---: | :---: | :--- |
+| 🥇 **#1** | **Claude 3.7 Sonnet (Thinking)** | Anthropic (2026) | Frontier Hybrid | 🔒 Closed | 88.0% | **70.3%** | 1,400 t | State-of-the-art hybrid reasoning & adaptive SWE-bench |
+| 🥈 **#2** | **⚡ Simplicio 27B** | simpletibr (2026) | **27B Hybrid (DeltaNet)** | 🟢 **Open** | **96.5%** | **46.5%** | **480 t** *(⚡ -43% economy)* | **#1 in Atomic Surgical Diff & 50-Point Loop Protocol** |
+| 🥉 **#3** | **OpenAI o3-mini (High Effort)** | OpenAI (2026) | Frontier Reasoning | 🔒 Closed | 82.0% | 53.0% | 1,650 t | Deep algorithmic reasoning and strict logic verification |
+| **#4** | **OpenAI o3 (Full)** | OpenAI (2026) | Frontier Reasoning | 🔒 Closed | 83.5% | 55.4% | 2,200 t | Frontier heavyweight reasoning with deep branch search |
+| **#5** | **DeepSeek-R1 (Full 671B)** | DeepSeek (2026) | 671B MoE CoT | 🟢 Open | 76.5% | 49.2% | 1,850 t | Unsupervised open reasoning trained via pure RL |
+| **#6** | **DeepSeek-V3** | DeepSeek (2026) | 671B MoE (37B active) | 🟢 Open | 75.0% | 43.4% | 790 t | High knowledge density and exceptional cost profile |
+| **#7** | **Gemini 2.0 Pro (Experimental)** | Google (2026) | Frontier MoE | 🔒 Closed | 77.0% | 47.5% | 1,100 t | Deep multimodal reasoning across 2M token context |
+| **#8** | **Gemini 2.0 Flash (Thinking)** | Google (2026) | Frontier Flash | 🔒 Closed | 74.5% | 42.0% | 920 t | Ultra-low latency and multimodal tool calling |
+| **#9** | **GPT-4o (2026 Edition)** | OpenAI (2026) | Frontier Dense | 🔒 Closed | 73.5% | 38.8% | 720 t | Generalist stability, tool routing, and API maturity |
+| **#10** | **Qwen 2.5 Coder 32B** | Alibaba (2026) | 32B Dense | 🟢 Open | 73.0% | 39.8% | 740 t | Best dense pre-trained base model for code completion |
+| **#11** | **Llama 3.3 70B Instruct** | Meta (2026) | 70B Dense | 🟢 Open | 67.5% | 37.5% | 790 t | Meta's flagship open-weights enterprise baseline |
+| **#12** | **Qwen3.8-27B (Thinking)** | Alibaba (2026) | 27B Hybrid DeltaNet | 🟢 Open | 69.2% | 40.8% | 850 t | Native DeltaNet hybrid reasoning with verbose CoT |
+| **#13** | **Codestral 25.01** | Mistral (2026) | 24B Dense | 🟢 Open | 68.0% | 33.0% | 660 t | Mistral's latest release focused on IDEs and FIM |
+| **#14** | **Qwen3.8-27B (Fast)** | Alibaba (2026) | 27B Hybrid DeltaNet | 🟢 Open | 58.4% | 33.5% | 590 t | Ultra-fast generation prone to diff formatting errors |
+| **#15** | **Yi-Coder 9B** | 01.AI (2026) | 9B Dense | 🟢 Open | 52.0% | 23.5% | 650 t | Lightweight model for resource-constrained environments |
 ---