Instructions to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp 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 AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16
# Run inference directly in the terminal:
llama cli -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16
# Run inference directly in the terminal:
llama cli -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp: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 AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16
# Run inference directly in the terminal:
./build/bin/llama cli -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16

Use Docker

docker model run hf.co/AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16

Llama new
LM Studio
Jan
Ollama
How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with Ollama:
```
ollama run hf.co/AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16
```
Unsloth Desktop

How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp: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": "AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with Docker Model Runner:
```
docker model run hf.co/AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16
```

Lemonade

How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16

Run and chat with the model

lemonade run user.Qwen3.5-9B-AX8860-llama.cpp-BF16

List all available models

lemonade list

Hermes Agent

How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp: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 AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp: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 "AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp: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"

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Qwen3.5-9B on AX8860 (llama.cpp + ggml-axera, 2026-11-05)

在爱芯 AX8860 的 NPU（4 核）上用 llama.cpp 运行 Qwen3.5-9B，支持文本和图像输入；一套 kernel 覆盖到 256k 上下文，NPU 内存随 llama.cpp 的上下文长度（-c）分配。 Qwen3.5-9B with text and image input on the 4-core NPU of the Axera AX8860, with llama.cpp; one kernel set for contexts up to 256k, the NPU memory allocated for the context llama.cpp is given (-c).

文件 / Files

文件 / File	说明 / Description
`qwen3.5-9b-mxfp4.gguf`	模型权重（GGUF MXFP4，6.00 GB，md5 `57b4b4b8fe3e7d05af29019438e5369a`）/ weights
`mmproj-qwen3.5-9b-bf16.gguf`	视觉编码器权重（BF16，921 MB，md5 `69d1a645084d0f31d4296a82dce991cf`）/ vision encoder weights
`bin/`	llama.cpp 板端程序（静态链接，只依赖系统 `/opt/lib` 与 libc）/ the board programs (static; need only the system `/opt/lib` and libc)
`kernels/`	NPU kernel + `manifest.json`：prefill、16 层 decode group（两组）、输出头、3072-patch ViT / the NPU kernels + `manifest.json`: prefill, a 16-layer decode group (two groups), the output head, the 3072-patch ViT
`run.sh`	运行脚本 / run script

权重只来自 GGUF：kernel 不含权重，加载时从 GGUF 转换到 NPU（约 36 - 38 s）；NPU 内存里只放一份权重（prefill 和 decode 共用）。 The weights come only from the GGUFs: the kernels hold none; they are converted from the GGUF at load (~36 - 38 s); one copy of the weights in NPU memory (shared by prefill and decode).

运行 / Run

环境 / Requirements：AX8860，SDK V0.5.0 或更新（cat /proc/ax_proc/version）/ SDK V0.5.0 or newer.

./run.sh "The capital of France is" 64                 # 文本补全 / text completion
./run.sh -i bus.jpg "Describe this image." 256         # 图像描述 / describe an image
./run.sh -c                                            # 命令行多轮对话 / multi-turn chat in the terminal
./run.sh -s 8080                                       # OpenAI 兼容服务 / OpenAI-compatible server
CTX=262144 ./run.sh -s 8080                            # 256k 上下文 / the full 256k context
CTX=16384 ./run.sh -s 8080                             # 16k 上下文，NPU 内存最省 / 16k context, the least NPU memory
curl http://<board>:8080/v1/chat/completions -H 'Content-Type: application/json' \
  -d '{"messages":[{"role":"user","content":"你好"}]}'

CTX 是传给 llama.cpp 的 -c（默认 131072，最大 262144）：KV cache 的 NPU 内存按它分配，所以短上下文的场景把它调小即可省内存。 CTX is llama.cpp's -c (default 131072, up to 262144): the NPU memory of the KV cache is allocated for it, so a smaller value saves memory for short-context use.

直接调用程序时设置 / When calling the programs directly: LD_LIBRARY_PATH=/opt/lib GGML_AXERA_KERNEL_DIR=<repo>/kernels，-ngl 99，-c 不超过 262144，服务端 -np 1。GGML_AXERA_PROF=1 退出时打印各阶段耗时 / prints per-stage times at exit。

性能 / Performance

AX8860，SDK V0.5.1，NPU 4 核 / 4 cores，贪心解码 / greedy decoding：

项 / Item	值 / Value
加载后 NPU 内存（CMM）/ NPU memory after load	`CTX=4096` 6.0 GB，`CTX=16384` 6.5 GB，`CTX=131072`（默认 / default）9.7 GB，`CTX=262144` 13.9 GB
加载后系统内存 / host memory after load	0.4（峰值 / peak 1.3） GB
加载 / load	36 - 38 s
4k prompt prefill	2898 tokens/s
短上下文 decode / short-context decode	42.5 tokens/s
32k prompt prefill / decode（32k 上下文 / context）	2458 tokens/s / 30.3 tokens/s
64k prompt prefill / decode（64k 上下文 / context）	1990 tokens/s / 23.3 tokens/s
120k prompt（113,463 token）prefill / decode（113k 上下文 / context）	1481 tokens/s / 17.7 tokens/s
图像编码 / image encoding（3072 patch = 768 图像 token / image tokens）	0.28 s

上下文越长 decode 越慢（attention 要读整个 KV）/ decode slows with the context (attention reads the whole KV).

正确性 / Correctness

"The capital of France is" → "Paris. …"；图像描述正确（车辆、文字、人物）/ image descriptions correct.
长文 PPL（-c 4096，英文小说）：5.014 / long-text PPL (-c 4096, an English novel): 5.014；中文 PPL（内部测试集，-c 512）：20.83，同一 GGUF 在 CPU 上 20.55 / Chinese PPL on an internal set: 20.83 (the same GGUF on the CPU: 20.55).
大海捞针 / needle in a haystack（一句暗号藏在小说 30% 处，末尾提问 / a code hidden at 30% of a novel, asked at the end）：32k / 64k / 120k token 全部命中 / found at 32k / 64k / 120k tokens.
KV cache 按 -c 分配后的输出与分配整个 256k 时一致（GGML_AXERA_KV_PAGES=1024 对照）/ the output with the KV cache allocated for -c matches the one with the whole 256k allocated.
同一输入多次运行，结果偶有极小差异（NPU 上累加顺序不固定），不影响输出质量。 Repeated runs of one input may differ slightly (the NPU's accumulation order varies); the output quality is the same.

已知限制 / Known limits

运行时独占 NPU 4 核，同时运行其他 NPU 程序会出错 / takes all 4 NPU cores; do not run other NPU programs alongside.
单序列：服务端须 -np 1 / one sequence: the server needs -np 1.
板子 16 GB 系统内存：llama-perplexity 的 -c 不超过 4096（它把整段上下文的 logits 放在系统内存里）/ with the board's 16 GB host memory, llama-perplexity works up to -c 4096 (it keeps the whole context's logits in host memory).
上下文越长 decode 越慢（attention 要读整个 KV）：64k 上下文 23.3 tokens/s，113k 17.7 tokens/s / decode slows with the context (attention reads the whole KV): 23.3 tokens/s at 64k, 17.7 at 113k.

源码 / Sources

llama.cpp：ZHEQIUSHUI/llama.cpp axera-dev @ 57a30bd（ggml/src/ggml-axera/models/qwen35_whole）
kernel：ZHEQIUSHUI/ax-dsl-kernels main @ f267e04（models/qwen35_9b/whole，kernels.md5 列出本套件 / lists this set）

Downloads last month: 220

GGUF

Model size

9B params

Architecture

qwen35

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Collection including AXERA-TECH/Qwen3.5-9B-AX8860-llama.cpp

AX8860

Collection

14 items • Updated 1 day ago