Instructions to use SC117/Agents-A1-MTP-APEX-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SC117/Agents-A1-MTP-APEX-GGUF with Transformers:

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

pipe = pipeline("text-generation", model="SC117/Agents-A1-MTP-APEX-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("SC117/Agents-A1-MTP-APEX-GGUF", dtype="auto")

llama-cpp-python

How to use SC117/Agents-A1-MTP-APEX-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="SC117/Agents-A1-MTP-APEX-GGUF",
	filename="Agents-A1-MTP-APEX-I-Balanced.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 SC117/Agents-A1-MTP-APEX-GGUF 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 SC117/Agents-A1-MTP-APEX-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf SC117/Agents-A1-MTP-APEX-GGUF:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf SC117/Agents-A1-MTP-APEX-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf SC117/Agents-A1-MTP-APEX-GGUF: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 SC117/Agents-A1-MTP-APEX-GGUF:BF16
# Run inference directly in the terminal:
./llama-cli -hf SC117/Agents-A1-MTP-APEX-GGUF: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 SC117/Agents-A1-MTP-APEX-GGUF:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf SC117/Agents-A1-MTP-APEX-GGUF:BF16

Use Docker

docker model run hf.co/SC117/Agents-A1-MTP-APEX-GGUF:BF16

LM Studio
Jan

vLLM

How to use SC117/Agents-A1-MTP-APEX-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SC117/Agents-A1-MTP-APEX-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SC117/Agents-A1-MTP-APEX-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SC117/Agents-A1-MTP-APEX-GGUF:BF16

SGLang

How to use SC117/Agents-A1-MTP-APEX-GGUF 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 "SC117/Agents-A1-MTP-APEX-GGUF" \
    --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": "SC117/Agents-A1-MTP-APEX-GGUF",
		"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 "SC117/Agents-A1-MTP-APEX-GGUF" \
        --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": "SC117/Agents-A1-MTP-APEX-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use SC117/Agents-A1-MTP-APEX-GGUF with Ollama:
```
ollama run hf.co/SC117/Agents-A1-MTP-APEX-GGUF:BF16
```

Unsloth Studio

How to use SC117/Agents-A1-MTP-APEX-GGUF 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 SC117/Agents-A1-MTP-APEX-GGUF 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 SC117/Agents-A1-MTP-APEX-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for SC117/Agents-A1-MTP-APEX-GGUF to start chatting

How to use SC117/Agents-A1-MTP-APEX-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SC117/Agents-A1-MTP-APEX-GGUF:BF16

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": "SC117/Agents-A1-MTP-APEX-GGUF:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use SC117/Agents-A1-MTP-APEX-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SC117/Agents-A1-MTP-APEX-GGUF: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 SC117/Agents-A1-MTP-APEX-GGUF:BF16

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use SC117/Agents-A1-MTP-APEX-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SC117/Agents-A1-MTP-APEX-GGUF: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 "SC117/Agents-A1-MTP-APEX-GGUF: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"

Docker Model Runner
How to use SC117/Agents-A1-MTP-APEX-GGUF with Docker Model Runner:
```
docker model run hf.co/SC117/Agents-A1-MTP-APEX-GGUF:BF16
```

Lemonade

How to use SC117/Agents-A1-MTP-APEX-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull SC117/Agents-A1-MTP-APEX-GGUF:BF16

Run and chat with the model

lemonade run user.Agents-A1-MTP-APEX-GGUF-BF16

List all available models

lemonade list

Agents-A1-MTP-APEX-GGUF / README_zh.md

SC117

Upload 2 files

e93eb5c verified about 14 hours ago

preview code

Raw

History Blame

24.5 kB

	---
	library_name: transformers
	license: apache-2.0
	license_link: https://huggingface.co/InternScience/Agents-A1/blob/main/LICENSE
	pipeline_tag: text-generation
	tags:
	- qwen3_5_moe
	- qwen3_5
	- reasoning
	- agentic
	- mtp
	- apex
	- quantization
	- gguf
	- multimodal
	base_model:
	- InternScience/Agents-A1
	---
	<div style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin-bottom: 24px;">
	<div style="background: #f5f5f7; border-radius: 20px; padding: 36px 32px; margin-bottom: 20px; text-align: center; position: relative; overflow: hidden;">
	<div style="position: absolute; top: -30px; right: -30px; width: 120px; height: 120px; background: #bbf7d0; border-radius: 50%;"></div>
	<div style="position: absolute; bottom: -20px; left: 40px; width: 80px; height: 80px; background: #86efac; border-radius: 50%;"></div>
	<div style="position: absolute; top: 50%; left: -15px; width: 60px; height: 60px; background: #bbf7d0; border-radius: 50%;"></div>
	<div style="display: inline-flex; gap: 8px; margin-bottom: 16px; position: relative; z-index: 1;">
	<span style="background: #10b981; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">APEX</span>
	<span style="background: #6366f1; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">MTP</span>
	<span style="background: #0ea5e9; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">Vision</span>
	<span style="background: #f97316; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">Apache-2.0</span>
	</div>
	<h1 style="margin: 0 0 8px 0; font-size: 32px; font-weight: 700; color: #064e3b; letter-spacing: -0.5px; border: none; position: relative; z-index: 1;">Agents-A1-MTP-APEX</h1>
	<p style="margin: 8px 0 0 0; font-size: 14px; position: relative; z-index: 1;"><a href="https://huggingface.co/SC117/Agents-A1-MTP-APEX-GGUF/blob/main/README.md" style="color: #10b981; text-decoration: none;">English</a> \| <span style="color: #6b7280;">📖 中文文档</span></p>
	<p style="margin: 0; font-size: 15px; color: #6b7280; position: relative; z-index: 1;">35B Agentic MoE 达到万亿参数性能 · APEX 量化 GGUF + BF16 + mmproj</p>
	</div>
	</div>

	<div style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; display: flex; flex-direction: column; gap: 20px; margin-bottom: 30px;">

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🤖</span> 关于 Agents-A1</div>
	<div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;">
	<p style="margin: 0 0 12px 0;"><a href="https://huggingface.co/InternScience/Agents-A1" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">Agents-A1</a> 是 <a href="https://huggingface.co/InternScience" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">InternScience</a> 提出的 <b>35B 参数 Mixture-of-Experts Agentic 模型</b>，在 <a href="https://huggingface.co/Qwen/Qwen3.5-35B-A3B" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">Qwen3.5-35B-A3B</a> 的基础上通过三阶段范式后训练得到：全域 SFT → 域级教师训练 → 多教师多域 on-policy 蒸馏。</p>
	<p style="margin: 0 0 12px 0;">尽管参数规模只有 ~35B，Agents-A1 在与 GPT-5.5、DeepSeek-V4-pro、Kimi-K2.6 等前沿大模型的对比中依然表现强劲——在 Seal-0 (56.4)、HiPhO (46.4)、FrontierScience-Olympiad (79.0)、IFBench (80.6)、IFEval (94.8) 达到 SOTA，同时在 BrowseComp (75.5)、XBench-DS-2510 (86.0)、GAIA (96.0)、SciCode (44.3)、HLE (47.6)、MolBench-bind (56.8) 等任务上于同尺寸模型中排名第一。</p>
	<p style="margin: 0;">本 GGUF 包包含 <b>mmproj-F16.gguf</b> 视觉投影器（支持 llama.cpp 多模态）。MTP 层从 <a href="https://huggingface.co/Qwen/Qwen3.5-35B-A3B" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">Qwen3.5-35B-A3B</a> 中提取并注入到 Agents-A1 的 safetensors（详见 <b>MTP 提取与注入</b> 章节）。<b>许可证：Apache-2.0。</b></p>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🧠</span> 模型详情</div>
	<div style="padding: 16px;">
	<table style="width: 100%; border-collapse: collapse; font-size: 13px;"><tbody>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">架构</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">Qwen3.5 MoE（Mixture of Experts）</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">参数规模</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">35B 总参数，每 token 激活 3B</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">专家数</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">256 个路由专家，每 token 激活 8 个</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">层数</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">40 层 Transformer + 1 层 MTP</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">上下文</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">262,144 tokens</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">MTP 来源</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">Qwen3.5-35B-A3B（1 层，785 个 tensor，注入）</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">Block Count</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white;">41（blk.0–39 + blk.40 MTP）</td></tr>
	<tr><td style="padding: 6px 10px; font-weight: bold; color: #334155; background: white;">许可证</td><td style="padding: 6px 10px; color: #4b5563; background: white;">Apache-2.0</td></tr>
	</tbody></table>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🔧</span> MTP 提取与注入</div>
	<div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;">
	<p style="margin: 0 0 12px 0;">官方发布的 <a href="https://huggingface.co/InternScience/Agents-A1" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">InternScience/Agents-A1</a> checkpoint 是一个 <b>40 层 Qwen3.5-35B-A3B MoE</b>，不包含 MTP（Multi-Token Prediction）层。为了在 llama.cpp 中启用 MTP 加速（长上下文生成提速 10–30%），我们 <b>从 Qwen3.5-35B-A3B 中提取 1 层 MTP</b>，注入到 Agents-A1 的 safetensors 中，再转 GGUF。</p>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">步骤 1 — 从 Qwen3.5-35B-A3B 提取 MTP tensor</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre;">源：J:\Models\Qwen3.5-35B-A3B-MTP（Qwen3.5-35B-A3B + 原生 MTP）
	from safetensors import safe_open
	import json, os
	·
	src = r"J:\Models\Qwen3.5-35B-A3B-MTP"
	with open(os.path.join(src, "model.safetensors.index.json")) as f:
	idx = json.load(f)
	mtp_keys = [k for k in idx["weight_map"] if "mtp" in k.lower()]
	print(f"Found {len(mtp_keys)} MTP tensors") # 785</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">步骤 2 — 作为新分片（N+1）追加</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre;">把 785 个 MTP tensor 保存为新分片
	new_shard = "model.safetensors-15-of-15.safetensors"
	save_file({k: get_tensor(k) for k in mtp_keys}, new_shard)
	·
	更新 model.safetensors.index.json：
	· metadata.total_size += 新分片大小
	· weight_map: 为每个 MTP key 追加新分片路径
	· 不修改原 14 个分片（避免触碰原始数据）</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">步骤 3 — 用 master llama.cpp 转 BF16 GGUF</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre;">F:\llama.cpp\llama.cpp-master\convert_hf_to_gguf.py ^
	J:\Models\Agents-A1 ^
	--outfile J:\Models\Agents-A1-MTP-GGUF\Agents-A1-MTP-BF16.gguf ^
	--outtype f16
	·
	master 版本自动处理 Qwen3.5MoE + MTP：
	· 常规层：blk.0–39
	· MTP 层：blk.40.nextn.* （785 个 tensor）</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">步骤 4 — 用 APEX 量化（Q4_K_M 默认，MTP 用 Q8_0）</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre;">F:\llama.cpp\...\llama-quantize.exe ^
	--imatrix J:\Models\Qwen3.5-35B-A3B.imatrix.gguf ^
	--tensor-type-file E:\apex-quant\configs\qwen36_35b_mtp_<档位>.txt ^
	J:\Models\Agents-A1-MTP-GGUF\Agents-A1-MTP-BF16.gguf ^
	J:\Models\Agents-A1-MTP-GGUF\Agents-A1-MTP-APEX-I-<档位>.gguf ^
	Q4_K_M
	·
	APEX qwen36_35b_mtp_*.txt 配置已包含 blk.40 override
	（所有档位 MTP 用 Q8_0）—— 无需手动 patch。</pre>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>📊</span> BenchLocal 测试结果（APEX-I-Compact, 16.14 GB）</div>
	<div style="padding: 16px;">
	<table style="width: 100%; border-collapse: collapse; font-size: 13px;"><thead><tr style="background: rgba(16,185,129,0.05);"><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">模式</th><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">ToolCall-15</th><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">BugFind-15</th><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">HermesAgent-20</th><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">能力上限</th><th style="padding: 7px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold;">实用得分</th></tr></thead><tbody>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">思考</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">100</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">88</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">87</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #047857; background: white;">91.2</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">71.2</td></tr>
	<tr><td style="padding: 6px 10px; font-weight: bold; color: #334155; background: white;">无思考</td><td style="padding: 6px 10px; color: #334155; background: white;">97</td><td style="padding: 6px 10px; color: #334155; background: white;">100</td><td style="padding: 6px 10px; color: #334155; background: white;">85</td><td style="padding: 6px 10px; font-weight: bold; color: #047857; background: white;">93.1</td><td style="padding: 6px 10px; color: #334155; background: white;">57.1</td></tr>
	</tbody></table>
	<p style="margin: 12px 0 0 0; font-size: 12px; color: #64748b; font-style: italic;">测试环境：RTX 5070 Ti 16GB + 128GB RAM · 无思考模式能力上限更高（BugFind +12），但复杂 agent 场景重试更多。</p>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🚀</span> 使用方法</div>
	<div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;">
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">llama.cpp（纯文本）</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre-wrap;">hf download SC117/Agents-A1-MTP-APEX-GGUF --include "*.gguf" --local-dir ./models
	./llama-server -m ./models/Agents-A1-MTP-APEX-I-Compact.gguf -ngl 99 -c 131072</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">llama.cpp（视觉 + 文本）</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre-wrap;">./llama-server -m ./models/Agents-A1-MTP-APEX-I-Compact.gguf --mmproj ./models/mmproj-F16.gguf -ngl 99 -c 131072</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">vLLM</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre-wrap;">vllm serve SC117/Agents-A1-MTP-APEX-GGUF --port 8000 --tensor-parallel-size 1 --max-model-len 262144 --reasoning-parser qwen3
	·
	工具调用变体
	vllm serve SC117/Agents-A1-MTP-APEX-GGUF --port 8000 --tensor-parallel-size 1 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder</pre>
	<p style="margin: 0 0 8px 0; font-weight: bold; color: #064e3b;">SGLang</p>
	<pre style="margin: 0; font-family: monospace; background: #f8fafc; padding: 10px 14px; border-radius: 6px; border: 1px solid #e2e8f0; font-size: 12px; color: #1e293b; white-space: pre-wrap;">python3 -m sglang.launch_server --model-path "SC117/Agents-A1-MTP-APEX-GGUF" --host 0.0.0.0 --port 30000</pre>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🎛️</span> 推荐采样参数</div>
	<div style="padding: 16px;">
	<p style="margin: 0 0 12px 0; font-size: 13px; color: #334155;">摘自 <a href="https://huggingface.co/InternScience/Agents-A1#recommended-sampling-parameters" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">官方 Agents-A1 模型卡</a>：</p>
	<table style="width: 100%; border-collapse: collapse; font-size: 13px;"><tbody>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">temperature</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white; font-family: monospace;">0.85</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">top_p</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white; font-family: monospace;">0.95</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">top_k</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white; font-family: monospace;">20</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">min_p</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white; font-family: monospace;">0.0</td></tr>
	<tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155; background: white;">presence_penalty</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563; background: white; font-family: monospace;">1.1</td></tr>
	<tr><td style="padding: 6px 10px; font-weight: bold; color: #334155; background: white;">repetition_penalty</td><td style="padding: 6px 10px; color: #4b5563; font-family: monospace; background: white;">1.0</td></tr>
	</tbody></table>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>💡</span> 什么是 APEX？</div>
	<div style="padding: 16px;">
	<p style="margin: 0 0 12px 0; font-size: 13px; color: #334155; line-height: 1.7;">本仓库的 GGUF 文件使用 <a href="https://github.com/mudler/apex-quant" target="_blank" style="color: #047857; text-decoration: none; font-weight: 700;">APEX</a> 量化——一种 MoE 感知的混合精度量化技术。APEX 对每个 tensor 按角色分类（路由专家、共享专家、SSM、注意力），按层级应用精度梯度：边缘敏感层（含 MTP）更高精度，中间冗余层更激进压缩。</p>
	<p style="margin: 0; font-size: 13px; color: #334155; line-height: 1.7; font-weight: 700;">APEX 在文件大小减半的前提下，PPL 仍优于 Q8_0，在部分场景甚至优于 F16。</p>
	<p style="margin: 12px 0 0 0; font-size: 13px; color: #334155;"><code>qwen36_35b_mtp_*.txt</code> 配置已包含 <b>blk.40</b>（MTP 层）的 override，所有四个 I 档位均以 Q8_0 保留。imatrix <code>Qwen3.5-35B-A3B.imatrix.gguf</code> 直接复用（同架构、专家布局兼容）。</p>
	</div>
	</div>

	<div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);">
	<div style="background: linear-gradient(135deg, #10b981 0%, #059669 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>📦</span> APEX 量化档位</div>
	<div style="padding: 16px;">
	<table style="width: 100%; border-collapse: collapse; font-size: 13px;"><thead><tr style="background: rgba(16,185,129,0.05);"><th style="padding: 8px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold; width: 35%;">文件</th><th style="padding: 8px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold; width: 15%;">大小</th><th style="padding: 8px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold; width: 15%;">档位</th><th style="padding: 8px 10px; border-bottom: 2px solid #10b981; text-align: left; color: #047857; font-weight: bold; width: 35%;">适用场景</th></tr></thead><tbody>
	<tr><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;"><code>*-APEX-I-Quality.gguf</code></td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">21.75 GB</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">I-Quality</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">高质量（Q6_K + iq4_xs 注意力）</td></tr>
	<tr><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;"><code>*-APEX-I-Balanced.gguf</code></td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">24.21 GB</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">I-Balanced</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #334155; background: white;">全能型（Q6_K + Q5_K 专家）</td></tr>
	<tr><td style="padding: 8px 10px; color: #334155;"><code>*-APEX-I-Compact.gguf</code></td><td style="padding: 8px 10px; color: #334155;">16.14 GB</td><td style="padding: 8px 10px; color: #334155;">I-Compact</td><td style="padding: 8px 10px; color: #334155;">质量/体积最佳平衡（Q4_K 默认）</td></tr>
	<tr><td style="padding: 8px 10px; color: #334155;"><code>*-APEX-I-Mini.gguf</code></td><td style="padding: 8px 10px; color: #334155;">13.36 GB</td><td style="padding: 8px 10px; color: #334155;">I-Mini</td><td style="padding: 8px 10px; color: #334155;">极致压缩，适配 16GB 显存（Q3_K + iq2_s）</td></tr>
	</tbody></table>
	<p style="margin: 12px 0 0 0; font-size: 12px; color: #64748b; font-style: italic;">BF16 源文件：<code>Agents-A1-MTP-BF16.gguf</code>（66.19 GB）。imatrix：<code>Qwen3.5-35B-A3B.imatrix.gguf</code>（从基座模型复用）。</p>
	</div>
	</div>

	</div>

	## 相关链接

	- 原始模型：https://huggingface.co/InternScience/Agents-A1
	- 基座模型（MTP 来源）：https://huggingface.co/Qwen/Qwen3.5-35B-A3B
	- 论文：https://arxiv.org/abs/2606.30616
	- APEX 量化：https://github.com/mudler/apex-quant
	- BenchLocal 测试：https://scorp1o117.github.io/benchlocal-results/

	## 引用

	```bibtex
	@misc{bai2026scalinghorizonparametersreaching,
	title={Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent},
	author={Lei Bai and Zongsheng Cao and Yang Chen and Zhiyao Cui and Shangheng Du and Yue Fan and Shiyang Feng and Zijie Guo and Haonan He and Liang He and Xiaohan He and Shuyue Hu and Yusong Hu and Songtao Huang and Yichen Jiang and Hao Li and Xin Li and Dahua Lin and Weihao Lin and Fenghua Ling and Dongrui Liu and Zhuo Liu and Runmin Ma and Chunjiang Mu and others},
	year={2026},
	eprint={2606.30616},
	archivePrefix={arXiv}
	}
	```