{ "@context": "https://schema.org", "@id": "https://aggr.aymericbeaumet.com/items/qwen-ai/2026-04-22-qwen3-6-27b-flagship-level-coding-in-a-27b-dense-model/#webpage", "@type": [ "WebPage", "ArchiveComponent" ], "articleBody": "![Qwen3.6-27B Main Image](https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.6/Figures/3.6_27b_banner.png)\n\n[QWEN STUDIO](https://chat.qwen.ai) [HUGGING FACE](https://huggingface.co/Qwen/Qwen3.6-27B) [MODELSCOPE](https://modelscope.cn/models/Qwen/Qwen3.6-27B) [DISCORD](https://discord.gg/yPEP2vHTu4)\n\nFollowing the launch of [Qwen3.6-Plus](https://qwen.ai/blog?id=qwen3.6) and [Qwen3.6-35B-A3B](https://qwen.ai/blog?id=qwen3.6-35b-a3b), we are excited to open-source **Qwen3.6-27B** — a dense 27-billion-parameter multimodal model at the scale the community has been asking for most. Still supporting both multimodal thinking and non-thinking modes, Qwen3.6-27B delivers flagship-level agentic coding performance, **surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B** (397B total / 17B active MoE) across all major coding benchmarks. As a dense architecture, it is straightforward to deploy without MoE routing complexity, making it an ideal choice for developers who need top-tier coding capabilities at a practical, widely-deployable scale. Qwen3.6-27B is now live on Qwen Studio, available through our API, and released as open weights for the community.\n\n- **Qwen3.6-27B** is a fully open-source dense model (27B parameters), featuring:\n - flagship-level agentic coding that surpasses Qwen3.5-397B-A17B\n - strong text and multimodal reasoning ability\n- You can chat interactively on [Qwen Studio](https://chat.qwen.ai), call via API on [Alibaba Cloud Model Studio API](https://modelstudio.alibabacloud.com/) (coming soon), or download weights from [Hugging Face](https://huggingface.co/Qwen/Qwen3.6-27B) and [ModelScope](https://modelscope.cn/models/Qwen/Qwen3.6-27B).\n\n![](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3.6/Figures/qwen3.6_27b_score.png)\n\n## Performance\n\nBelow we present comprehensive evaluations of Qwen3.6-27B against both dense and MoE baselines, including our previous-generation open-source flagship Qwen3.5-397B-A17B. Qwen3.6-27B delivers remarkable improvements across agentic coding benchmarks, surpassing models with up to 15x its total parameter count.\n\n### Language\n\nQwen3.6-27B achieves a breakthrough in agentic coding for dense models. With only 27B parameters, it outperforms the Qwen3.5-397B-A17B (397B total / 17B active) on every major coding benchmark — including SWE-bench Verified (77.2 vs. 76.2), SWE-bench Pro (53.5 vs. 50.9), Terminal-Bench 2.0 (59.3 vs. 52.5), and SkillsBench (48.2 vs. 30.0). It also surpasses all peer-scale dense models by a wide margin. On reasoning tasks, Qwen3.6-27B achieves 87.8 on GPQA Diamond, competitive with models several times its size.\n\n| | Qwen3.5-27B | Qwen3.5-397B-A17B | Gemma4-31B | Claude 4.5 Opus | Qwen3.6-35B-A3B | Qwen3.6-27B |\n| ---------------------- | ----------- | ----------------- | ---------- | --------------- | --------------- | ----------- |\n| Coding Agent | | | | | | |\n| SWE-bench Verified | 75.0 | 76.2 | 52.0 | 80.9 | 73.4 | 77.2 |\n| SWE-bench Pro | 51.2 | 50.9 | 35.7 | 57.1 | 49.5 | 53.5 |\n| SWE-bench Multilingual | 69.3 | 69.3 | 51.7 | 77.5 | 67.2 | 71.3 |\n| Terminal-Bench 2.0 | 41.6 | 52.5 | 42.9 | 59.3 | 51.5 | 59.3 |\n| SkillsBench Avg5 | 27.2 | 30.0 | 23.6 | 45.3 | 28.7 | 48.2 |\n| QwenWebBench | 1068 | 1186 | 1197 | 1536 | 1397 | 1487 |\n| NL2Repo | 27.3 | 32.2 | 15.5 | 43.2 | 29.4 | 36.2 |\n| Claw-Eval Avg | 64.3 | 70.7 | 48.5 | 76.6 | 68.7 | 72.4 |\n| Claw-Eval Pass^3 | 46.2 | 48.1 | 25.0 | 59.6 | 50.0 | 60.6 |\n| QwenClawBench | 52.2 | 51.8 | 41.7 | 52.3 | 52.6 | 53.4 |\n| Knowledge | | | | | | |\n| MMLU-Pro | 86.1 | 87.8 | 85.2 | 89.5 | 85.2 | 86.2 |\n| MMLU-Redux | 93.2 | 94.9 | 93.7 | 95.6 | 93.3 | 93.5 |\n| SuperGPQA | 65.6 | 70.4 | 65.7 | 70.6 | 64.7 | 66.0 |\n| C-Eval | 90.5 | 93.0 | 82.6 | 92.2 | 90.0 | 91.4 |\n| STEM & Reasoning | | | | | | |\n| GPQA Diamond | 85.5 | 88.4 | 84.3 | 87.0 | 86.0 | 87.8 |\n| HLE | 24.3 | 28.7 | 19.5 | 30.8 | 21.4 | 24.0 |\n| LiveCodeBench v6 | 80.7 | 83.6 | 80.0 | 84.8 | 80.4 | 83.9 |\n| HMMT Feb 25 | 92.0 | 94.8 | 88.7 | 92.9 | 90.7 | 93.8 |\n| HMMT Nov 25 | 89.8 | 92.7 | 87.5 | 93.3 | 89.1 | 90.7 |\n| HMMT Feb 26 | 84.3 | 87.9 | 77.2 | 85.3 | 83.6 | 84.3 |\n| IMOAnswerBench | 79.9 | 80.9 | 74.5 | 84.0 | 78.9 | 80.8 |\n| AIME26 | 92.6 | 93.3 | 89.2 | 95.1 | 92.7 | 94.1 |\n\n\\* SWE-Bench Series: Internal agent scaffold (bash + file-edit tools); temp=1.0, top\\_p=0.95, 200K context window. We correct some problematic tasks in the public set of SWE-bench Pro and evaluate all baselines on the refined benchmark.\\\n\\* Terminal-Bench 2.0: Harbor/Terminus-2 harness; 3h timeout, 32 CPU/48 GB RAM; temp=1.0, top\\_p=0.95, top\\_k=20, max\\_tokens=80K, 256K ctx; avg of 5 runs.\\\n\\* SkillsBench: Evaluated via OpenCode on 78 tasks (self-contained subset, excluding API-dependent tasks); avg of 5 runs.\\\n\\* NL2Repo: Others are evaluated via Claude Code (temp=1.0, top\\_p=0.95, max\\_turns=900).\\\n\\* QwenClawBench: A real-user-distribution Claw agent benchmark; temp=0.6, 256K ctx.\\\n\\* QwenWebBench: An internal front-end code generation benchmark; bilingual (EN/CN), 7 categories (Web Design, Web Apps, Games, SVG, Data Visualization, Animation, and 3D); auto-render + multimodal judge (code/visual correctness); BT/Elo rating system.\\\n\\* AIME 26: We use the full AIME 2026 (I & II), where the scores may differ from Qwen 3.5 notes.\n\n### Vision Language\n\nQwen3.6-27B is natively multimodal, supporting both vision-language thinking and non-thinking modes in a single unified checkpoint — the same as Qwen3.6-35B-A3B. It handles images and video alongside text, enabling multimodal reasoning, document understanding, and visual question answering.\n\n| | Qwen3.5-27B | Qwen3.5-397B-A17B | Gemma4-31B | Claude 4.5 Opus | Qwen3.6-35B-A3B | Qwen3.6-27B |\n| ---------------------- | ----------- | ----------------- | ---------- | --------------- | --------------- | ----------- |\n| STEM & Puzzle | | | | | | |\n| MMMU | 82.3 | 85.0 | 80.4 | 80.7 | 81.7 | 82.9 |\n| MMMU-Pro | 75.0 | 79.0 | 76.9 | 70.6 | 75.3 | 75.8 |\n| MathVista mini | 87.8 | -- | 79.3 | -- | 86.4 | 87.4 |\n| DynaMath | 87.7 | 86.3 | 79.5 | 79.7 | 82.8 | 85.6 |\n| VlmsAreBlind | 96.9 | -- | 87.2 | -- | 96.6 | 97.0 |\n| General VQA | | | | | | |\n| RealWorldQA | 83.7 | 83.9 | 72.3 | 77.0 | 85.3 | 84.1 |\n| MMStar | 81.0 | 83.8 | 77.3 | 73.2 | 80.7 | 81.4 |\n| MMBenchEN-DEV-v1.1 | 92.6 | -- | 90.9 | -- | 92.8 | 92.3 |\n| SimpleVQA | 56.0 | 67.1 | 52.9 | 65.7 | 58.9 | 56.1 |\n| Document Understanding | | | | | | |\n| CharXiv RQ | 79.5 | 80.8 | 67.9 | 68.5 | 78.0 | 78.4 |\n| CC-OCR | 81.0 | 82.0 | 75.7 | 76.9 | 81.9 | 81.2 |\n| OCRBench | 89.4 | -- | 86.1 | -- | 90.0 | 89.4 |\n| Spatial Intelligence | | | | | | |\n| ERQA | 60.5 | 67.5 | 57.5 | 46.8 | 61.8 | 62.5 |\n| CountBench | 97.8 | 97.2 | 96.1 | 90.6 | 96.1 | 97.8 |\n| RefCOCO avg | 90.9 | 92.3 | -- | -- | 92.0 | 92.5 |\n| EmbSpatialBench | 84.5 | -- | -- | -- | 84.3 | 84.6 |\n| RefSpatialBench | 67.7 | -- | 4.7 | -- | 64.3 | 70.0 |\n| Video Understanding | | | | | | |\n| VideoMME(w sub.) | 87.0 | 87.5 | -- | 77.7 | 86.6 | 87.7 |\n| VideoMMMU | 82.3 | 84.7 | 81.6 | 84.4 | 83.7 | 84.4 |\n| MLVU | 85.9 | 86.7 | -- | 81.7 | 86.2 | 86.6 |\n| MVBench | 74.6 | 77.6 | -- | 67.2 | 74.6 | 75.5 |\n| Visual Agent | | | | | | |\n| V\\* | 93.7 | 95.8 | -- | 67.0 | 90.1 | 94.7 |\n| AndroidWorld | 64.2 | -- | -- | -- | -- | 70.3 |\n\n\\* Empty cells (--) indicate scores not yet available or not applicable.\n\n## Build with Qwen3.6-27B\n\nQwen3.6-27B is coming soon to Alibaba Cloud Model Studio. Please stand by until we are fully ready.\n\nQwen3.6-27B is available as open weights on [Hugging Face](https://huggingface.co/Qwen/Qwen3.6-27B) and [ModelScope](https://modelscope.cn/models/Qwen/Qwen3.6-27B) for self-hosting, and through the [Alibaba Cloud Model Studio](https://modelstudio.alibabacloud.com/) API. You can also try it instantly on [Qwen Studio](https://chat.qwen.ai).\n\nThe model can be seamlessly integrated with popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code, to streamline development workflows and enable efficient, context-aware coding experiences.\n\n### API Usage\n\nThis release supports the `preserve_thinking` feature: preserving thinking content from all preceding turns in messages, which is **recommended for agentic tasks**.\n\n#### Alibaba Cloud Model Studio\n\nAlibaba Cloud Model Studio supports industry-standard protocols, including chat completions and responses APIs compatible with OpenAI’s specification, as well as an API interface compatible with Anthropic.\n\nExample code for chat completions API is provided below:\n\n```python\n\"\"\"\nEnvironment variables (per official docs):\n DASHSCOPE_API_KEY: Your API Key from https://modelstudio.console.alibabacloud.com\n DASHSCOPE_BASE_URL: (optional) Base URL for compatible-mode API.\n - Beijing: https://dashscope.aliyuncs.com/compatible-mode/v1\n - Singapore: https://dashscope-intl.aliyuncs.com/compatible-mode/v1\n - US (Virginia): https://dashscope-us.aliyuncs.com/compatible-mode/v1\n DASHSCOPE_MODEL: (optional) Model name; override for different models.\n\"\"\"\nfrom openai import OpenAI\nimport os\n\napi_key = os.environ.get(\"DASHSCOPE_API_KEY\")\nif not api_key:\n raise ValueError(\n \"DASHSCOPE_API_KEY is required. \"\n \"Set it via: export DASHSCOPE_API_KEY='your-api-key'\"\n )\n\nclient = OpenAI(\n api_key=api_key,\n base_url=os.environ.get(\n \"DASHSCOPE_BASE_URL\",\n \"https://dashscope-intl.aliyuncs.com/compatible-mode/v1\",\n ),\n)\n\nmessages = [{\"role\": \"user\", \"content\": \"Introduce vibe coding.\"}]\n\nmodel = os.environ.get(\n \"DASHSCOPE_MODEL\",\n \"qwen3.6-27b\",\n)\ncompletion = client.chat.completions.create(\n model=model,\n messages=messages,\n extra_body={\n \"enable_thinking\": True,\n # \"preserve_thinking\": True,\n },\n stream=True\n)\n\nreasoning_content = \"\" # Full reasoning trace\nanswer_content = \"\" # Full response\nis_answering = False # Whether we have entered the answer phase\nprint(\"\\n\" + \"=\" * 20 + \"Reasoning\" + \"=\" * 20 + \"\\n\")\n\nfor chunk in completion:\n if not chunk.choices:\n print(\"\\nUsage:\")\n print(chunk.usage)\n continue\n\n delta = chunk.choices[0].delta\n\n # Collect reasoning content only\n if hasattr(delta, \"reasoning_content\") and delta.reasoning_content is not None:\n if not is_answering:\n print(delta.reasoning_content, end=\"\", flush=True)\n reasoning_content += delta.reasoning_content\n\n # Received content, start answer phase\n if hasattr(delta, \"content\") and delta.content:\n if not is_answering:\n print(\"\\n\" + \"=\" * 20 + \"Answer\" + \"=\" * 20 + \"\\n\")\n is_answering = True\n print(delta.content, end=\"\", flush=True)\n answer_content += delta.content\n```\n\nFor more information, please visit the [API doc](https://modelstudio.console.alibabacloud.com/?tab=doc#/doc/?type=model&url=2840915).\n\n### Coding & Agents\n\nQwen3.6-27B features excellent agentic coding capabilities and can be seamlessly integrated into popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code.\n\n#### OpenClaw\n\nQwen3.6-27B is compatible with [OpenClaw](https://openclaw.ai) (formerly Moltbot / Clawdbot), a self-hosted open-source AI coding agent. Connect it to [Model Studio](https://www.alibabacloud.com/help/en/model-studio/openclaw) to get a full agentic coding experience in the terminal. Get started with the following script:\n\n```bash\n# Node.js 22+\ncurl -fsSL https://molt.bot/install.sh | bash # macOS / Linux\n\n# Set your API key\nexport DASHSCOPE_API_KEY=\n\n# Launch OpenClaw\nopenclaw dashboard # web browser\n# openclaw tui # Open a new terminal and start the TUI\n```\n\nOn first use, edit `~/.openclaw/openclaw.json` to point OpenClaw at Model Studio. Find or create the following fields and merge them — **do not overwrite the entire file** to preserve your existing settings:\n\n```json\n{\n \"models\": {\n \"mode\": \"merge\",\n \"providers\": {\n \"modelstudio\": {\n \"baseUrl\": \"https://dashscope-intl.aliyuncs.com/compatible-mode/v1\",\n \"apiKey\": \"DASHSCOPE_API_KEY\",\n \"api\": \"openai-completions\",\n \"models\": [\n {\n \"id\": \"qwen3.6-27b\",\n \"name\": \"qwen3.6-27b\",\n \"reasoning\": true,\n \"input\": [\"text\", \"image\"],\n \"contextWindow\": 131072,\n \"maxTokens\": 16384\n }\n ]\n }\n }\n },\n \"agents\": {\n \"defaults\": {\n \"model\": {\n \"primary\": \"modelstudio/qwen3.6-27b\"\n },\n \"models\": {\n \"modelstudio/qwen3.6-27b\": {}\n }\n }\n }\n}\n```\n\n#### Qwen Code\n\nQwen3.6-27B is compatible with [Qwen Code](https://qwen.ai/qwencode), an open-source AI agent designed for the terminal and deeply optimized for the Qwen Series. Get started with the following script:\n\n```bash\n# Node.js 20+\nnpm install -g @qwen-code/qwen-code@latest\n\n# Start Qwen Code (interactive)\nqwen\n\n# Then, in the session:\n/help\n/auth\n```\n\nOn first use, you’ll be prompted to sign in. You can run `/auth` anytime to switch authentication methods.\n\n#### Claude Code\n\nQwen APIs also support the Anthropic API protocol, meaning you can use it with tools like **Claude Code** for elevated coding experience:\n\n```bash\n# Install Claude Code\nnpm install -g @anthropic-ai/claude-code\n\n# Configure environment\nexport ANTHROPIC_MODEL=\"qwen3.6-27b\"\nexport ANTHROPIC_SMALL_FAST_MODEL=\"qwen3.6-27b\"\nexport ANTHROPIC_BASE_URL=https://dashscope-intl.aliyuncs.com/apps/anthropic\nexport ANTHROPIC_AUTH_TOKEN=\n\n# Launch the CLI\nclaude\n```\n\n## Summary\n\nQwen3.6-27B demonstrates that a well-trained dense model can surpass much larger predecessors on the tasks that matter most for developers. At 27 billion parameters — the most widely deployed open-source scale — it outperforms the 397B-parameter Qwen3.5-397B-A17B on every major agentic coding benchmark, while remaining straightforward to deploy and serve. With Qwen3.6-27B joining the roster, the Qwen3.6 open-source family now offers a comprehensive range of models, underscoring a generation where agentic coding achieved breakthroughs across every scale — from the 3B-active Qwen3.6-35B-A3B to the API-accessible Qwen3.6-Plus and Qwen3.6-Max-Preview. We are grateful for the community’s feedback and look forward to seeing what you build with these models. Stay tuned for more from the Qwen team!\n\n## Citation\n\nFeel free to cite the following article if you find Qwen3.6-27B helpful:\n\n```bibtex\n@misc{qwen36_27b,\n title = {{Qwen3.6-27B}: Flagship-Level Coding in a 27B Dense Model},\n url = {https://qwen.ai/blog?id=qwen3.6-27b},\n author = {{Qwen Team}},\n month = {April},\n year = {2026}\n}\n```\n", "authors": [ "QwenTeam" ], "captured_at": "2026-10-19T01:26:53.998049939+00:00", "category": "ai", "content_html": "

\"Qwen3.6-27B

\n

QWEN STUDIO HUGGING FACE MODELSCOPE DISCORD

\n

Following the launch of Qwen3.6-Plus and Qwen3.6-35B-A3B, we are excited to open-source Qwen3.6-27B — a dense 27-billion-parameter multimodal model at the scale the community has been asking for most. Still supporting both multimodal thinking and non-thinking modes, Qwen3.6-27B delivers flagship-level agentic coding performance, surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B (397B total / 17B active MoE) across all major coding benchmarks. As a dense architecture, it is straightforward to deploy without MoE routing complexity, making it an ideal choice for developers who need top-tier coding capabilities at a practical, widely-deployable scale. Qwen3.6-27B is now live on Qwen Studio, available through our API, and released as open weights for the community.

\n\n

\"\"

\n

Performance

\n

Below we present comprehensive evaluations of Qwen3.6-27B against both dense and MoE baselines, including our previous-generation open-source flagship Qwen3.5-397B-A17B. Qwen3.6-27B delivers remarkable improvements across agentic coding benchmarks, surpassing models with up to 15x its total parameter count.

\n

Language

\n

Qwen3.6-27B achieves a breakthrough in agentic coding for dense models. With only 27B parameters, it outperforms the Qwen3.5-397B-A17B (397B total / 17B active) on every major coding benchmark — including SWE-bench Verified (77.2 vs. 76.2), SWE-bench Pro (53.5 vs. 50.9), Terminal-Bench 2.0 (59.3 vs. 52.5), and SkillsBench (48.2 vs. 30.0). It also surpasses all peer-scale dense models by a wide margin. On reasoning tasks, Qwen3.6-27B achieves 87.8 on GPQA Diamond, competitive with models several times its size.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Qwen3.5-27BQwen3.5-397B-A17BGemma4-31BClaude 4.5 OpusQwen3.6-35B-A3BQwen3.6-27B
Coding Agent
SWE-bench Verified75.076.252.080.973.477.2
SWE-bench Pro51.250.935.757.149.553.5
SWE-bench Multilingual69.369.351.777.567.271.3
Terminal-Bench 2.041.652.542.959.351.559.3
SkillsBench Avg527.230.023.645.328.748.2
QwenWebBench106811861197153613971487
NL2Repo27.332.215.543.229.436.2
Claw-Eval Avg64.370.748.576.668.772.4
Claw-Eval Pass^346.248.125.059.650.060.6
QwenClawBench52.251.841.752.352.653.4
Knowledge
MMLU-Pro86.187.885.289.585.286.2
MMLU-Redux93.294.993.795.693.393.5
SuperGPQA65.670.465.770.664.766.0
C-Eval90.593.082.692.290.091.4
STEM & Reasoning
GPQA Diamond85.588.484.387.086.087.8
HLE24.328.719.530.821.424.0
LiveCodeBench v680.783.680.084.880.483.9
HMMT Feb 2592.094.888.792.990.793.8
HMMT Nov 2589.892.787.593.389.190.7
HMMT Feb 2684.387.977.285.383.684.3
IMOAnswerBench79.980.974.584.078.980.8
AIME2692.693.389.295.192.794.1
\n

* SWE-Bench Series: Internal agent scaffold (bash + file-edit tools); temp=1.0, top_p=0.95, 200K context window. We correct some problematic tasks in the public set of SWE-bench Pro and evaluate all baselines on the refined benchmark.
\n* Terminal-Bench 2.0: Harbor/Terminus-2 harness; 3h timeout, 32 CPU/48 GB RAM; temp=1.0, top_p=0.95, top_k=20, max_tokens=80K, 256K ctx; avg of 5 runs.
\n* SkillsBench: Evaluated via OpenCode on 78 tasks (self-contained subset, excluding API-dependent tasks); avg of 5 runs.
\n* NL2Repo: Others are evaluated via Claude Code (temp=1.0, top_p=0.95, max_turns=900).
\n* QwenClawBench: A real-user-distribution Claw agent benchmark; temp=0.6, 256K ctx.
\n* QwenWebBench: An internal front-end code generation benchmark; bilingual (EN/CN), 7 categories (Web Design, Web Apps, Games, SVG, Data Visualization, Animation, and 3D); auto-render + multimodal judge (code/visual correctness); BT/Elo rating system.
\n* AIME 26: We use the full AIME 2026 (I & II), where the scores may differ from Qwen 3.5 notes.

\n

Vision Language

\n

Qwen3.6-27B is natively multimodal, supporting both vision-language thinking and non-thinking modes in a single unified checkpoint — the same as Qwen3.6-35B-A3B. It handles images and video alongside text, enabling multimodal reasoning, document understanding, and visual question answering.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Qwen3.5-27BQwen3.5-397B-A17BGemma4-31BClaude 4.5 OpusQwen3.6-35B-A3BQwen3.6-27B
STEM & Puzzle
MMMU82.385.080.480.781.782.9
MMMU-Pro75.079.076.970.675.375.8
MathVista mini87.8--79.3--86.487.4
DynaMath87.786.379.579.782.885.6
VlmsAreBlind96.9--87.2--96.697.0
General VQA
RealWorldQA83.783.972.377.085.384.1
MMStar81.083.877.373.280.781.4
MMBenchEN-DEV-v1.192.6--90.9--92.892.3
SimpleVQA56.067.152.965.758.956.1
Document Understanding
CharXiv RQ79.580.867.968.578.078.4
CC-OCR81.082.075.776.981.981.2
OCRBench89.4--86.1--90.089.4
Spatial Intelligence
ERQA60.567.557.546.861.862.5
CountBench97.897.296.190.696.197.8
RefCOCO avg90.992.3----92.092.5
EmbSpatialBench84.5------84.384.6
RefSpatialBench67.7--4.7--64.370.0
Video Understanding
VideoMME(w sub.)87.087.5--77.786.687.7
VideoMMMU82.384.781.684.483.784.4
MLVU85.986.7--81.786.286.6
MVBench74.677.6--67.274.675.5
Visual Agent
V*93.795.8--67.090.194.7
AndroidWorld64.2--------70.3
\n

* Empty cells (--) indicate scores not yet available or not applicable.

\n

Build with Qwen3.6-27B

\n

Qwen3.6-27B is coming soon to Alibaba Cloud Model Studio. Please stand by until we are fully ready.

\n

Qwen3.6-27B is available as open weights on Hugging Face and ModelScope for self-hosting, and through the Alibaba Cloud Model Studio API. You can also try it instantly on Qwen Studio.

\n

The model can be seamlessly integrated with popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code, to streamline development workflows and enable efficient, context-aware coding experiences.

\n

API Usage

\n

This release supports the preserve_thinking feature: preserving thinking content from all preceding turns in messages, which is recommended for agentic tasks.

\n

Alibaba Cloud Model Studio

\n

Alibaba Cloud Model Studio supports industry-standard protocols, including chat completions and responses APIs compatible with OpenAI’s specification, as well as an API interface compatible with Anthropic.

\n

Example code for chat completions API is provided below:

\n
"""\nEnvironment variables (per official docs):\n  DASHSCOPE_API_KEY: Your API Key from https://modelstudio.console.alibabacloud.com\n  DASHSCOPE_BASE_URL: (optional) Base URL for compatible-mode API.\n    - Beijing: https://dashscope.aliyuncs.com/compatible-mode/v1\n    - Singapore: https://dashscope-intl.aliyuncs.com/compatible-mode/v1\n    - US (Virginia): https://dashscope-us.aliyuncs.com/compatible-mode/v1\n  DASHSCOPE_MODEL: (optional) Model name; override for different models.\n"""\nfrom openai import OpenAI\nimport os\n\napi_key = os.environ.get("DASHSCOPE_API_KEY")\nif not api_key:\n    raise ValueError(\n        "DASHSCOPE_API_KEY is required. "\n        "Set it via: export DASHSCOPE_API_KEY='your-api-key'"\n    )\n\nclient = OpenAI(\n    api_key=api_key,\n    base_url=os.environ.get(\n        "DASHSCOPE_BASE_URL",\n        "https://dashscope-intl.aliyuncs.com/compatible-mode/v1",\n    ),\n)\n\nmessages = [{"role": "user", "content": "Introduce vibe coding."}]\n\nmodel = os.environ.get(\n    "DASHSCOPE_MODEL",\n    "qwen3.6-27b",\n)\ncompletion = client.chat.completions.create(\n    model=model,\n    messages=messages,\n    extra_body={\n        "enable_thinking": True,\n        # "preserve_thinking": True,\n    },\n    stream=True\n)\n\nreasoning_content = ""  # Full reasoning trace\nanswer_content = ""  # Full response\nis_answering = False  # Whether we have entered the answer phase\nprint("\\n" + "=" * 20 + "Reasoning" + "=" * 20 + "\\n")\n\nfor chunk in completion:\n    if not chunk.choices:\n        print("\\nUsage:")\n        print(chunk.usage)\n        continue\n\n    delta = chunk.choices[0].delta\n\n    # Collect reasoning content only\n    if hasattr(delta, "reasoning_content") and delta.reasoning_content is not None:\n        if not is_answering:\n            print(delta.reasoning_content, end="", flush=True)\n        reasoning_content += delta.reasoning_content\n\n    # Received content, start answer phase\n    if hasattr(delta, "content") and delta.content:\n        if not is_answering:\n            print("\\n" + "=" * 20 + "Answer" + "=" * 20 + "\\n")\n            is_answering = True\n        print(delta.content, end="", flush=True)\n        answer_content += delta.content\n
\n

For more information, please visit the API doc.

\n

Coding & Agents

\n

Qwen3.6-27B features excellent agentic coding capabilities and can be seamlessly integrated into popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code.

\n

OpenClaw

\n

Qwen3.6-27B is compatible with OpenClaw (formerly Moltbot / Clawdbot), a self-hosted open-source AI coding agent. Connect it to Model Studio to get a full agentic coding experience in the terminal. Get started with the following script:

\n
# Node.js 22+\ncurl -fsSL https://molt.bot/install.sh | bash   # macOS / Linux\n\n# Set your API key\nexport DASHSCOPE_API_KEY=<your_api_key>\n\n# Launch OpenClaw\nopenclaw dashboard # web browser\n# openclaw tui # Open a new terminal and start the TUI\n
\n

On first use, edit ~/.openclaw/openclaw.json to point OpenClaw at Model Studio. Find or create the following fields and merge them — do not overwrite the entire file to preserve your existing settings:

\n
{\n  "models": {\n    "mode": "merge",\n    "providers": {\n      "modelstudio": {\n        "baseUrl": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1",\n        "apiKey": "DASHSCOPE_API_KEY",\n        "api": "openai-completions",\n        "models": [\n          {\n            "id": "qwen3.6-27b",\n            "name": "qwen3.6-27b",\n            "reasoning": true,\n            "input": ["text", "image"],\n            "contextWindow": 131072,\n            "maxTokens": 16384\n          }\n        ]\n      }\n    }\n  },\n  "agents": {\n    "defaults": {\n      "model": {\n        "primary": "modelstudio/qwen3.6-27b"\n      },\n      "models": {\n        "modelstudio/qwen3.6-27b": {}\n      }\n    }\n  }\n}\n
\n

Qwen Code

\n

Qwen3.6-27B is compatible with Qwen Code, an open-source AI agent designed for the terminal and deeply optimized for the Qwen Series. Get started with the following script:

\n
# Node.js 20+\nnpm install -g @qwen-code/qwen-code@latest\n\n# Start Qwen Code (interactive)\nqwen\n\n# Then, in the session:\n/help\n/auth\n
\n

On first use, you’ll be prompted to sign in. You can run /auth anytime to switch authentication methods.

\n

Claude Code

\n

Qwen APIs also support the Anthropic API protocol, meaning you can use it with tools like Claude Code for elevated coding experience:

\n
# Install Claude Code\nnpm install -g @anthropic-ai/claude-code\n\n# Configure environment\nexport ANTHROPIC_MODEL="qwen3.6-27b"\nexport ANTHROPIC_SMALL_FAST_MODEL="qwen3.6-27b"\nexport ANTHROPIC_BASE_URL=https://dashscope-intl.aliyuncs.com/apps/anthropic\nexport ANTHROPIC_AUTH_TOKEN=<your_api_key>\n\n# Launch the CLI\nclaude\n
\n

Summary

\n

Qwen3.6-27B demonstrates that a well-trained dense model can surpass much larger predecessors on the tasks that matter most for developers. At 27 billion parameters — the most widely deployed open-source scale — it outperforms the 397B-parameter Qwen3.5-397B-A17B on every major agentic coding benchmark, while remaining straightforward to deploy and serve. With Qwen3.6-27B joining the roster, the Qwen3.6 open-source family now offers a comprehensive range of models, underscoring a generation where agentic coding achieved breakthroughs across every scale — from the 3B-active Qwen3.6-35B-A3B to the API-accessible Qwen3.6-Plus and Qwen3.6-Max-Preview. We are grateful for the community’s feedback and look forward to seeing what you build with these models. Stay tuned for more from the Qwen team!

\n

Citation

\n

Feel free to cite the following article if you find Qwen3.6-27B helpful:

\n
@misc{qwen36_27b,\n    title = {{Qwen3.6-27B}: Flagship-Level Coding in a 27B Dense Model},\n    url = {https://qwen.ai/blog?id=qwen3.6-27b},\n    author = {{Qwen Team}},\n    month = {April},\n    year = {2026}\n}\n
\n", "content_markdown": "![Qwen3.6-27B Main Image](https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.6/Figures/3.6_27b_banner.png)\n\n[QWEN STUDIO](https://chat.qwen.ai) [HUGGING FACE](https://huggingface.co/Qwen/Qwen3.6-27B) [MODELSCOPE](https://modelscope.cn/models/Qwen/Qwen3.6-27B) [DISCORD](https://discord.gg/yPEP2vHTu4)\n\nFollowing the launch of [Qwen3.6-Plus](https://qwen.ai/blog?id=qwen3.6) and [Qwen3.6-35B-A3B](https://qwen.ai/blog?id=qwen3.6-35b-a3b), we are excited to open-source **Qwen3.6-27B** — a dense 27-billion-parameter multimodal model at the scale the community has been asking for most. Still supporting both multimodal thinking and non-thinking modes, Qwen3.6-27B delivers flagship-level agentic coding performance, **surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B** (397B total / 17B active MoE) across all major coding benchmarks. As a dense architecture, it is straightforward to deploy without MoE routing complexity, making it an ideal choice for developers who need top-tier coding capabilities at a practical, widely-deployable scale. Qwen3.6-27B is now live on Qwen Studio, available through our API, and released as open weights for the community.\n\n- **Qwen3.6-27B** is a fully open-source dense model (27B parameters), featuring:\n - flagship-level agentic coding that surpasses Qwen3.5-397B-A17B\n - strong text and multimodal reasoning ability\n- You can chat interactively on [Qwen Studio](https://chat.qwen.ai), call via API on [Alibaba Cloud Model Studio API](https://modelstudio.alibabacloud.com/) (coming soon), or download weights from [Hugging Face](https://huggingface.co/Qwen/Qwen3.6-27B) and [ModelScope](https://modelscope.cn/models/Qwen/Qwen3.6-27B).\n\n![](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3.6/Figures/qwen3.6_27b_score.png)\n\n## Performance\n\nBelow we present comprehensive evaluations of Qwen3.6-27B against both dense and MoE baselines, including our previous-generation open-source flagship Qwen3.5-397B-A17B. Qwen3.6-27B delivers remarkable improvements across agentic coding benchmarks, surpassing models with up to 15x its total parameter count.\n\n### Language\n\nQwen3.6-27B achieves a breakthrough in agentic coding for dense models. With only 27B parameters, it outperforms the Qwen3.5-397B-A17B (397B total / 17B active) on every major coding benchmark — including SWE-bench Verified (77.2 vs. 76.2), SWE-bench Pro (53.5 vs. 50.9), Terminal-Bench 2.0 (59.3 vs. 52.5), and SkillsBench (48.2 vs. 30.0). It also surpasses all peer-scale dense models by a wide margin. On reasoning tasks, Qwen3.6-27B achieves 87.8 on GPQA Diamond, competitive with models several times its size.\n\n| | Qwen3.5-27B | Qwen3.5-397B-A17B | Gemma4-31B | Claude 4.5 Opus | Qwen3.6-35B-A3B | Qwen3.6-27B |\n| ---------------------- | ----------- | ----------------- | ---------- | --------------- | --------------- | ----------- |\n| Coding Agent | | | | | | |\n| SWE-bench Verified | 75.0 | 76.2 | 52.0 | 80.9 | 73.4 | 77.2 |\n| SWE-bench Pro | 51.2 | 50.9 | 35.7 | 57.1 | 49.5 | 53.5 |\n| SWE-bench Multilingual | 69.3 | 69.3 | 51.7 | 77.5 | 67.2 | 71.3 |\n| Terminal-Bench 2.0 | 41.6 | 52.5 | 42.9 | 59.3 | 51.5 | 59.3 |\n| SkillsBench Avg5 | 27.2 | 30.0 | 23.6 | 45.3 | 28.7 | 48.2 |\n| QwenWebBench | 1068 | 1186 | 1197 | 1536 | 1397 | 1487 |\n| NL2Repo | 27.3 | 32.2 | 15.5 | 43.2 | 29.4 | 36.2 |\n| Claw-Eval Avg | 64.3 | 70.7 | 48.5 | 76.6 | 68.7 | 72.4 |\n| Claw-Eval Pass^3 | 46.2 | 48.1 | 25.0 | 59.6 | 50.0 | 60.6 |\n| QwenClawBench | 52.2 | 51.8 | 41.7 | 52.3 | 52.6 | 53.4 |\n| Knowledge | | | | | | |\n| MMLU-Pro | 86.1 | 87.8 | 85.2 | 89.5 | 85.2 | 86.2 |\n| MMLU-Redux | 93.2 | 94.9 | 93.7 | 95.6 | 93.3 | 93.5 |\n| SuperGPQA | 65.6 | 70.4 | 65.7 | 70.6 | 64.7 | 66.0 |\n| C-Eval | 90.5 | 93.0 | 82.6 | 92.2 | 90.0 | 91.4 |\n| STEM & Reasoning | | | | | | |\n| GPQA Diamond | 85.5 | 88.4 | 84.3 | 87.0 | 86.0 | 87.8 |\n| HLE | 24.3 | 28.7 | 19.5 | 30.8 | 21.4 | 24.0 |\n| LiveCodeBench v6 | 80.7 | 83.6 | 80.0 | 84.8 | 80.4 | 83.9 |\n| HMMT Feb 25 | 92.0 | 94.8 | 88.7 | 92.9 | 90.7 | 93.8 |\n| HMMT Nov 25 | 89.8 | 92.7 | 87.5 | 93.3 | 89.1 | 90.7 |\n| HMMT Feb 26 | 84.3 | 87.9 | 77.2 | 85.3 | 83.6 | 84.3 |\n| IMOAnswerBench | 79.9 | 80.9 | 74.5 | 84.0 | 78.9 | 80.8 |\n| AIME26 | 92.6 | 93.3 | 89.2 | 95.1 | 92.7 | 94.1 |\n\n\\* SWE-Bench Series: Internal agent scaffold (bash + file-edit tools); temp=1.0, top\\_p=0.95, 200K context window. We correct some problematic tasks in the public set of SWE-bench Pro and evaluate all baselines on the refined benchmark.\\\n\\* Terminal-Bench 2.0: Harbor/Terminus-2 harness; 3h timeout, 32 CPU/48 GB RAM; temp=1.0, top\\_p=0.95, top\\_k=20, max\\_tokens=80K, 256K ctx; avg of 5 runs.\\\n\\* SkillsBench: Evaluated via OpenCode on 78 tasks (self-contained subset, excluding API-dependent tasks); avg of 5 runs.\\\n\\* NL2Repo: Others are evaluated via Claude Code (temp=1.0, top\\_p=0.95, max\\_turns=900).\\\n\\* QwenClawBench: A real-user-distribution Claw agent benchmark; temp=0.6, 256K ctx.\\\n\\* QwenWebBench: An internal front-end code generation benchmark; bilingual (EN/CN), 7 categories (Web Design, Web Apps, Games, SVG, Data Visualization, Animation, and 3D); auto-render + multimodal judge (code/visual correctness); BT/Elo rating system.\\\n\\* AIME 26: We use the full AIME 2026 (I & II), where the scores may differ from Qwen 3.5 notes.\n\n### Vision Language\n\nQwen3.6-27B is natively multimodal, supporting both vision-language thinking and non-thinking modes in a single unified checkpoint — the same as Qwen3.6-35B-A3B. It handles images and video alongside text, enabling multimodal reasoning, document understanding, and visual question answering.\n\n| | Qwen3.5-27B | Qwen3.5-397B-A17B | Gemma4-31B | Claude 4.5 Opus | Qwen3.6-35B-A3B | Qwen3.6-27B |\n| ---------------------- | ----------- | ----------------- | ---------- | --------------- | --------------- | ----------- |\n| STEM & Puzzle | | | | | | |\n| MMMU | 82.3 | 85.0 | 80.4 | 80.7 | 81.7 | 82.9 |\n| MMMU-Pro | 75.0 | 79.0 | 76.9 | 70.6 | 75.3 | 75.8 |\n| MathVista mini | 87.8 | -- | 79.3 | -- | 86.4 | 87.4 |\n| DynaMath | 87.7 | 86.3 | 79.5 | 79.7 | 82.8 | 85.6 |\n| VlmsAreBlind | 96.9 | -- | 87.2 | -- | 96.6 | 97.0 |\n| General VQA | | | | | | |\n| RealWorldQA | 83.7 | 83.9 | 72.3 | 77.0 | 85.3 | 84.1 |\n| MMStar | 81.0 | 83.8 | 77.3 | 73.2 | 80.7 | 81.4 |\n| MMBenchEN-DEV-v1.1 | 92.6 | -- | 90.9 | -- | 92.8 | 92.3 |\n| SimpleVQA | 56.0 | 67.1 | 52.9 | 65.7 | 58.9 | 56.1 |\n| Document Understanding | | | | | | |\n| CharXiv RQ | 79.5 | 80.8 | 67.9 | 68.5 | 78.0 | 78.4 |\n| CC-OCR | 81.0 | 82.0 | 75.7 | 76.9 | 81.9 | 81.2 |\n| OCRBench | 89.4 | -- | 86.1 | -- | 90.0 | 89.4 |\n| Spatial Intelligence | | | | | | |\n| ERQA | 60.5 | 67.5 | 57.5 | 46.8 | 61.8 | 62.5 |\n| CountBench | 97.8 | 97.2 | 96.1 | 90.6 | 96.1 | 97.8 |\n| RefCOCO avg | 90.9 | 92.3 | -- | -- | 92.0 | 92.5 |\n| EmbSpatialBench | 84.5 | -- | -- | -- | 84.3 | 84.6 |\n| RefSpatialBench | 67.7 | -- | 4.7 | -- | 64.3 | 70.0 |\n| Video Understanding | | | | | | |\n| VideoMME(w sub.) | 87.0 | 87.5 | -- | 77.7 | 86.6 | 87.7 |\n| VideoMMMU | 82.3 | 84.7 | 81.6 | 84.4 | 83.7 | 84.4 |\n| MLVU | 85.9 | 86.7 | -- | 81.7 | 86.2 | 86.6 |\n| MVBench | 74.6 | 77.6 | -- | 67.2 | 74.6 | 75.5 |\n| Visual Agent | | | | | | |\n| V\\* | 93.7 | 95.8 | -- | 67.0 | 90.1 | 94.7 |\n| AndroidWorld | 64.2 | -- | -- | -- | -- | 70.3 |\n\n\\* Empty cells (--) indicate scores not yet available or not applicable.\n\n## Build with Qwen3.6-27B\n\nQwen3.6-27B is coming soon to Alibaba Cloud Model Studio. Please stand by until we are fully ready.\n\nQwen3.6-27B is available as open weights on [Hugging Face](https://huggingface.co/Qwen/Qwen3.6-27B) and [ModelScope](https://modelscope.cn/models/Qwen/Qwen3.6-27B) for self-hosting, and through the [Alibaba Cloud Model Studio](https://modelstudio.alibabacloud.com/) API. You can also try it instantly on [Qwen Studio](https://chat.qwen.ai).\n\nThe model can be seamlessly integrated with popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code, to streamline development workflows and enable efficient, context-aware coding experiences.\n\n### API Usage\n\nThis release supports the `preserve_thinking` feature: preserving thinking content from all preceding turns in messages, which is **recommended for agentic tasks**.\n\n#### Alibaba Cloud Model Studio\n\nAlibaba Cloud Model Studio supports industry-standard protocols, including chat completions and responses APIs compatible with OpenAI’s specification, as well as an API interface compatible with Anthropic.\n\nExample code for chat completions API is provided below:\n\n```python\n\"\"\"\nEnvironment variables (per official docs):\n DASHSCOPE_API_KEY: Your API Key from https://modelstudio.console.alibabacloud.com\n DASHSCOPE_BASE_URL: (optional) Base URL for compatible-mode API.\n - Beijing: https://dashscope.aliyuncs.com/compatible-mode/v1\n - Singapore: https://dashscope-intl.aliyuncs.com/compatible-mode/v1\n - US (Virginia): https://dashscope-us.aliyuncs.com/compatible-mode/v1\n DASHSCOPE_MODEL: (optional) Model name; override for different models.\n\"\"\"\nfrom openai import OpenAI\nimport os\n\napi_key = os.environ.get(\"DASHSCOPE_API_KEY\")\nif not api_key:\n raise ValueError(\n \"DASHSCOPE_API_KEY is required. \"\n \"Set it via: export DASHSCOPE_API_KEY='your-api-key'\"\n )\n\nclient = OpenAI(\n api_key=api_key,\n base_url=os.environ.get(\n \"DASHSCOPE_BASE_URL\",\n \"https://dashscope-intl.aliyuncs.com/compatible-mode/v1\",\n ),\n)\n\nmessages = [{\"role\": \"user\", \"content\": \"Introduce vibe coding.\"}]\n\nmodel = os.environ.get(\n \"DASHSCOPE_MODEL\",\n \"qwen3.6-27b\",\n)\ncompletion = client.chat.completions.create(\n model=model,\n messages=messages,\n extra_body={\n \"enable_thinking\": True,\n # \"preserve_thinking\": True,\n },\n stream=True\n)\n\nreasoning_content = \"\" # Full reasoning trace\nanswer_content = \"\" # Full response\nis_answering = False # Whether we have entered the answer phase\nprint(\"\\n\" + \"=\" * 20 + \"Reasoning\" + \"=\" * 20 + \"\\n\")\n\nfor chunk in completion:\n if not chunk.choices:\n print(\"\\nUsage:\")\n print(chunk.usage)\n continue\n\n delta = chunk.choices[0].delta\n\n # Collect reasoning content only\n if hasattr(delta, \"reasoning_content\") and delta.reasoning_content is not None:\n if not is_answering:\n print(delta.reasoning_content, end=\"\", flush=True)\n reasoning_content += delta.reasoning_content\n\n # Received content, start answer phase\n if hasattr(delta, \"content\") and delta.content:\n if not is_answering:\n print(\"\\n\" + \"=\" * 20 + \"Answer\" + \"=\" * 20 + \"\\n\")\n is_answering = True\n print(delta.content, end=\"\", flush=True)\n answer_content += delta.content\n```\n\nFor more information, please visit the [API doc](https://modelstudio.console.alibabacloud.com/?tab=doc#/doc/?type=model&url=2840915).\n\n### Coding & Agents\n\nQwen3.6-27B features excellent agentic coding capabilities and can be seamlessly integrated into popular third-party coding assistants, including OpenClaw, Claude Code, and Qwen Code.\n\n#### OpenClaw\n\nQwen3.6-27B is compatible with [OpenClaw](https://openclaw.ai) (formerly Moltbot / Clawdbot), a self-hosted open-source AI coding agent. Connect it to [Model Studio](https://www.alibabacloud.com/help/en/model-studio/openclaw) to get a full agentic coding experience in the terminal. Get started with the following script:\n\n```bash\n# Node.js 22+\ncurl -fsSL https://molt.bot/install.sh | bash # macOS / Linux\n\n# Set your API key\nexport DASHSCOPE_API_KEY=\n\n# Launch OpenClaw\nopenclaw dashboard # web browser\n# openclaw tui # Open a new terminal and start the TUI\n```\n\nOn first use, edit `~/.openclaw/openclaw.json` to point OpenClaw at Model Studio. Find or create the following fields and merge them — **do not overwrite the entire file** to preserve your existing settings:\n\n```json\n{\n \"models\": {\n \"mode\": \"merge\",\n \"providers\": {\n \"modelstudio\": {\n \"baseUrl\": \"https://dashscope-intl.aliyuncs.com/compatible-mode/v1\",\n \"apiKey\": \"DASHSCOPE_API_KEY\",\n \"api\": \"openai-completions\",\n \"models\": [\n {\n \"id\": \"qwen3.6-27b\",\n \"name\": \"qwen3.6-27b\",\n \"reasoning\": true,\n \"input\": [\"text\", \"image\"],\n \"contextWindow\": 131072,\n \"maxTokens\": 16384\n }\n ]\n }\n }\n },\n \"agents\": {\n \"defaults\": {\n \"model\": {\n \"primary\": \"modelstudio/qwen3.6-27b\"\n },\n \"models\": {\n \"modelstudio/qwen3.6-27b\": {}\n }\n }\n }\n}\n```\n\n#### Qwen Code\n\nQwen3.6-27B is compatible with [Qwen Code](https://qwen.ai/qwencode), an open-source AI agent designed for the terminal and deeply optimized for the Qwen Series. Get started with the following script:\n\n```bash\n# Node.js 20+\nnpm install -g @qwen-code/qwen-code@latest\n\n# Start Qwen Code (interactive)\nqwen\n\n# Then, in the session:\n/help\n/auth\n```\n\nOn first use, you’ll be prompted to sign in. You can run `/auth` anytime to switch authentication methods.\n\n#### Claude Code\n\nQwen APIs also support the Anthropic API protocol, meaning you can use it with tools like **Claude Code** for elevated coding experience:\n\n```bash\n# Install Claude Code\nnpm install -g @anthropic-ai/claude-code\n\n# Configure environment\nexport ANTHROPIC_MODEL=\"qwen3.6-27b\"\nexport ANTHROPIC_SMALL_FAST_MODEL=\"qwen3.6-27b\"\nexport ANTHROPIC_BASE_URL=https://dashscope-intl.aliyuncs.com/apps/anthropic\nexport ANTHROPIC_AUTH_TOKEN=\n\n# Launch the CLI\nclaude\n```\n\n## Summary\n\nQwen3.6-27B demonstrates that a well-trained dense model can surpass much larger predecessors on the tasks that matter most for developers. At 27 billion parameters — the most widely deployed open-source scale — it outperforms the 397B-parameter Qwen3.5-397B-A17B on every major agentic coding benchmark, while remaining straightforward to deploy and serve. With Qwen3.6-27B joining the roster, the Qwen3.6 open-source family now offers a comprehensive range of models, underscoring a generation where agentic coding achieved breakthroughs across every scale — from the 3B-active Qwen3.6-35B-A3B to the API-accessible Qwen3.6-Plus and Qwen3.6-Max-Preview. We are grateful for the community’s feedback and look forward to seeing what you build with these models. Stay tuned for more from the Qwen team!\n\n## Citation\n\nFeel free to cite the following article if you find Qwen3.6-27B helpful:\n\n```bibtex\n@misc{qwen36_27b,\n title = {{Qwen3.6-27B}: Flagship-Level Coding in a 27B Dense Model},\n url = {https://qwen.ai/blog?id=qwen3.6-27b},\n author = {{Qwen Team}},\n month = {April},\n year = {2026}\n}\n```\n", "dateCreated": "2026-10-19T01:26:53.998049939+00:00", "date_created": "2026-04-22T02:00:00+00:00", "date_modified": null, "date_published": "2026-04-22T02:00:00+00:00", "description": "QWEN STUDIO HUGGING FACE MODELSCOPE DISCORD Following the launch of Qwen3.6-Plus and Qwen3.6-35B-A3B, we are excited to open-source Qwen3.6-27B — a dense 27-billion-parameter multimodal model at the scale the community has been asking for…", "external_url": "https://qwen.ai/blog?id=qwen3.6-27b", "git": { "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e", "path": "items/qwen-ai/2026/04/2026-04-22-qwen3-6-27b-flagship-level-coding-in-a-27b-dense-model.md", "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/qwen-ai/2026/04/2026-04-22-qwen3-6-27b-flagship-level-coding-in-a-27b-dense-model.md" }, "id": "urn:aggr:item:b903e8c984c5a892e87550badb859726465651f4", "isBasedOn": { "@id": "https://qwen.ai/blog?id=qwen3.6-27b" }, "mainEntity": { "@id": "https://qwen.ai/blog?id=qwen3.6-27b", "@type": "CreativeWork", "archivedAt": { "@id": "https://aggr.aymericbeaumet.com/items/qwen-ai/2026-04-22-qwen3-6-27b-flagship-level-coding-in-a-27b-dense-model/#webpage" }, "author": [ { "@type": "Person", "name": "QwenTeam" } ], "datePublished": "2026-04-22T02:00:00+00:00", "headline": "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model", "image": { "@type": "ImageObject", "caption": null, "height": 90, "url": "https://img.alicdn.com/imgextra/i2/O1CN01pnefgf1o0FcxKA3uu_!!6000000005162-2-tps-1590-954.png", "width": 160 }, "timeRequired": "PT8M", "url": "https://qwen.ai/blog?id=qwen3.6-27b", "wordCount": 1728 }, "name": "Archived snapshot: Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model", "preview": { "alt": null, "color": null, "height": 90, "placeholder": { "data_url": "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNk+A8AAQUBAScY42YAAAAASUVORK5CYII=", "hash": "" }, "url": "https://img.alicdn.com/imgextra/i2/O1CN01pnefgf1o0FcxKA3uu_!!6000000005162-2-tps-1590-954.png", "width": 160 }, "reading_minutes": 8, "replicated_at": null, "source": "qwen-ai", "source_name": "Qwen", "summary": null, "tags": [ "open-source" ], "temporalCoverage": "2026-10-19T01:26:53.998049939+00:00", "title": "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model", "url": "https://aggr.aymericbeaumet.com/items/qwen-ai/2026-04-22-qwen3-6-27b-flagship-level-coding-in-a-27b-dense-model/", "word_count": 1728 }