Instructions to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use JetBrains/Qwen3.8-3.6-27B-blend-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 JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

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 JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

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 JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

Use Docker

docker model run hf.co/JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "JetBrains/Qwen3.8-3.6-27B-blend-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": "JetBrains/Qwen3.8-3.6-27B-blend-GGUF",
		"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/JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

Ollama
How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with Ollama:
```
ollama run hf.co/JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
```
Unsloth Desktop

How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

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": "JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with Docker Model Runner:
```
docker model run hf.co/JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M
```

Lemonade

How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.8-3.6-27B-blend-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use JetBrains/Qwen3.8-3.6-27B-blend-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 JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

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 JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use JetBrains/Qwen3.8-3.6-27B-blend-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M

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 "JetBrains/Qwen3.8-3.6-27B-blend-GGUF:Q4_K_M" \
  --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"

vvsotnikov commited on 5 days ago

Commit

8afb890

verified ·

1 Parent(s): 7745fa5

Document unofficial derivative status, attribution, and modifications

Browse files

Files changed (3) hide show

CHANGES.md +37 -0
NOTICE +22 -0
README.md +11 -1

CHANGES.md ADDED Viewed

	@@ -0,0 +1,37 @@

+# Modifications to upstream Qwen material
+This distribution contains files modified by JetBrains from Qwen models developed by Alibaba Cloud. This document records the modifications and identifies the affected model files; it accompanies the binary artifacts listed below.
+## Base model merge
+The base model is a 50/50 combination of Qwen3.6-27B and Qwen3.8-27B, created by linearly interpolating their checkpoint parameters. The configuration, tokenizer, processor, and chat template are retained from Qwen3.8-27B. Exact source revisions and merge settings are recorded in [merge-manifest.json](merge-manifest.json).
+The interpolation uses float32 accumulation. Upstream sources:
+- [Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B/tree/6a9e13bd6fc8f0983b9b99948120bc37f49c13e9), revision `6a9e13bd6fc8f0983b9b99948120bc37f49c13e9`, weight 0.5.
+- [Qwen/Qwen3.8-27B](https://huggingface.co/Qwen/Qwen3.8-27B/tree/1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0), revision `1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0`, weight 0.5.
+## Distribution-specific changes
+The merged checkpoint was converted to GGUF and quantized to IQ3_S, Q4_K_M, and Q5_K_M from one BF16 intermediate, without an importance matrix. Each main GGUF retains the native MTP head. The vision encoder and projector were converted to a separate BF16 GGUF. Published filenames were updated to the JetBrains blend name.
+Exact conversion settings and runtime versions are recorded in [conversion-manifest.json](conversion-manifest.json). Historical source identifiers in the manifests are retained for reproducibility.
+## Modified model artifacts
+The following files contain the merged parameters and, where applicable, the format conversions described above. SHA-256 values identify the binaries to which this modification notice applies.
+| File | SHA-256 |
+|---|---|
+| `Qwen3.8-3.6-27B-blend-IQ3_S.gguf` | `94cc60e6ba6150003ae956d5a98c3914e8d9fe18ecafc3fe400e00232c8cecd2` |
+| `Qwen3.8-3.6-27B-blend-Q4_K_M.gguf` | `5f54ded85a78f11ad7a1c02d17564a268238ac8afe9b3d94fb0b614e473006a1` |
+| `Qwen3.8-3.6-27B-blend-Q5_K_M.gguf` | `d095b2807e5b120b57caf6a6c4bfeddf17f3ebf605279a64fe3a0fcb98722ba0` |
+| `mmproj-Qwen3.8-3.6-27B-blend-BF16.gguf` | `4568f7e9f79ef5b942be9acc8be717156482d0659a6fde111df660f57a1875c8` |
+## Documentation and attribution — 2026-09-15
+Model cards and repository references were adapted for the JetBrains distribution. The model cards now prominently identify this release as an unofficial JetBrains derivative and disclose the absence of Alibaba affiliation or endorsement. NOTICE and this modification record were added. This documentation update does not alter model artifacts, runtime configuration, or provenance manifests.
+The original Qwen LICENSE and Alibaba Cloud copyright notice are retained unchanged. Neither pinned upstream source contains a NOTICE file; this distribution's NOTICE was added by JetBrains.
+Copyright 2026 JetBrains N.V. applies only to original material contributed by JetBrains. Upstream Qwen material remains subject to its original copyright and attribution notices. See [LICENSE](LICENSE) and [NOTICE](NOTICE).

NOTICE ADDED Viewed

	@@ -0,0 +1,22 @@

+Qwen3.8-3.6-27B-blend-GGUF
+This distribution incorporates material from Qwen3.6-27B and
+Qwen3.8-27B, developed by Alibaba Cloud.
+Copyright 2026 Alibaba Cloud
+Original material contributed by JetBrains:
+Copyright 2026 JetBrains N.V.
+The JetBrains copyright notice applies only to original material
+contributed by JetBrains. It does not apply to upstream Qwen
+material or other third-party material.
+This is an unofficial derivative prepared by JetBrains. This release
+is not affiliated with, sponsored by, or endorsed by Alibaba Cloud
+or Alibaba Group. References to Qwen and Alibaba identify the
+upstream models and their origin.
+Distributed under the Apache License, Version 2.0.
+See LICENSE for the license terms and CHANGES.md for a description
+of the modifications.

README.md CHANGED Viewed

@@ -14,9 +14,11 @@ tags:
 # Qwen3.8-3.6-27B-blend GGUF
 IQ3_S, Q4_K_M, and Q5_K_M conversions of [JetBrains/Qwen3.8-3.6-27B-blend](https://huggingface.co/JetBrains/Qwen3.8-3.6-27B-blend/tree/f1a19acf58aa8caf7a0a507c9083245f79944dcc), pinned to revision `f1a19acf58aa8caf7a0a507c9083245f79944dcc`.
-The source is a 50/50 linear merge of Qwen3.6-27B and Qwen3.8-27B. Original merge provenance is in `merge-manifest.json`.
 | File | Size (decimal GB) | Contents |
 |---|---:|---|
@@ -71,3 +73,11 @@ All three passed local text, image, streaming, two-request concurrency, invalid-
 On the first arithmetic smoke request, Q4_K_M accepted 98/114 drafted tokens (86.0%) and Q5_K_M accepted 78/93 (83.9%). These are individual smoke results, not comparative quality or speed benchmarks. No BF16 parity or broad quality evaluation is claimed.
 IQ3_S was added on 2026-09-15 using the same BF16 intermediate and llama.cpp revision. It passed the same functional checks with thinking and native MTP enabled. Its arithmetic smoke request accepted 113/144 drafted tokens (78.5%). This is a smoke-test observation, not a quality or performance benchmark. The exact artifact SHA256, input SHA256, and validation results are recorded in `conversion-manifest.json`.

 # Qwen3.8-3.6-27B-blend GGUF
+**Unofficial JetBrains derivative.** This model was prepared by JetBrains using Qwen models developed by Alibaba Cloud. This release is not affiliated with, sponsored by, or endorsed by Alibaba Cloud or Alibaba Group.
 IQ3_S, Q4_K_M, and Q5_K_M conversions of [JetBrains/Qwen3.8-3.6-27B-blend](https://huggingface.co/JetBrains/Qwen3.8-3.6-27B-blend/tree/f1a19acf58aa8caf7a0a507c9083245f79944dcc), pinned to revision `f1a19acf58aa8caf7a0a507c9083245f79944dcc`.
+The base model is a 50/50 combination of Qwen3.6-27B and Qwen3.8-27B, created by linearly interpolating their checkpoint parameters. The configuration, tokenizer, processor, and chat template are retained from Qwen3.8-27B. Exact source revisions and merge settings are recorded in [merge-manifest.json](merge-manifest.json). See [conversion-manifest.json](conversion-manifest.json) for conversion settings and runtime versions.
 | File | Size (decimal GB) | Contents |
 |---|---:|---|
 On the first arithmetic smoke request, Q4_K_M accepted 98/114 drafted tokens (86.0%) and Q5_K_M accepted 78/93 (83.9%). These are individual smoke results, not comparative quality or speed benchmarks. No BF16 parity or broad quality evaluation is claimed.
 IQ3_S was added on 2026-09-15 using the same BF16 intermediate and llama.cpp revision. It passed the same functional checks with thinking and native MTP enabled. Its arithmetic smoke request accepted 113/144 drafted tokens (78.5%). This is a smoke-test observation, not a quality or performance benchmark. The exact artifact SHA256, input SHA256, and validation results are recorded in `conversion-manifest.json`.
+## License and attribution
+Distributed under the Apache License, Version 2.0. The original Qwen [LICENSE](LICENSE), including `Copyright 2026 Alibaba Cloud`, is retained unchanged.
+Copyright 2026 JetBrains N.V. This notice applies only to original material contributed by JetBrains. Upstream Qwen material remains subject to its original copyright and attribution notices.
+See [NOTICE](NOTICE) for attribution and [CHANGES.md](CHANGES.md) for a description of the modifications and the affected files.