diff --git a/.github/workflows/test-integration.yml b/.github/workflows/test-integration.yml
index c4354d1c..4f604db4 100644
--- a/.github/workflows/test-integration.yml
+++ b/.github/workflows/test-integration.yml
@@ -41,7 +41,7 @@ on:
       model:
         description: 'LLM model to use'
         required: false
-        default: 'qwen3.5-plus'
+        default: 'qwen3.6-plus'
 
 env:
   MANAGER_IMAGE: hiclaw/manager-agent:ci-test
@@ -105,7 +105,7 @@ jobs:
           HICLAW_LLM_API_KEY: ${{ secrets.HICLAW_LLM_API_KEY }}
         run: |
           RUNTIME="${{ inputs.worker_runtime || 'openclaw' }}"
-          MODEL="${{ inputs.model || 'qwen3.5-plus' }}"
+          MODEL="${{ inputs.model || 'qwen3.6-plus' }}"
           if [ "$RUNTIME" = "copaw" ]; then
             WORKER_IMG="${{ env.COPAW_WORKER_IMAGE }}"
             MANAGER_IMG="${{ env.MANAGER_COPAW_IMAGE }}"
@@ -369,7 +369,7 @@ jobs:
           echo "- Tests: \`$NON_GITHUB_TESTS\`" >> $GITHUB_STEP_SUMMARY
           echo "- Manager Image: \`${{ env.MANAGER_IMAGE }}\`" >> $GITHUB_STEP_SUMMARY
           echo "- Worker Runtime: \`${{ inputs.worker_runtime || 'openclaw' }}\`" >> $GITHUB_STEP_SUMMARY
-          echo "- Model: \`${{ inputs.model || 'qwen3.5-plus' }}\`" >> $GITHUB_STEP_SUMMARY
+          echo "- Model: \`${{ inputs.model || 'qwen3.6-plus' }}\`" >> $GITHUB_STEP_SUMMARY
 
 # ============================================================
 # Manual baseline re-generation for a specific release version
diff --git a/README.zh-CN.md b/README.zh-CN.md
index 771cf85e..f1c31690 100644
--- a/README.zh-CN.md
+++ b/README.zh-CN.md
@@ -76,7 +76,7 @@ Set-ExecutionPolicy Bypass -Scope Process -Force; $wc=New-Object Net.WebClient;
 
 第五步：选择模型接口。百炼 Coding Plan 和百炼通用接口有所不同，这里我们选择 Coding Plan 接口。[购买Coding Plan](https://bailian.console.aliyun.com/cn-beijing/?source_channel=4qjGAvs1Pl&tab=coding-plan#/efm/index)
 
-第六步：选择模型系列。如果第五步中选择的是百炼 Coding Plan，您可以选择 qwen3.5-plus、GLM等，待 Matrix room 建立起来后，还可通过发送指令，让 Manager 切换其他到模型。
+第六步：选择模型系列。如果第五步中选择的是百炼 Coding Plan，您可以选择 qwen3.6-plus、GLM等，待 Matrix room 建立起来后，还可通过发送指令，让 Manager 切换其他到模型。
 
 第七步：开始测试 API 联通性，若测试成功，效果如下。
 ![测试](https://img.alicdn.com/imgextra/i4/O1CN0148wFGG1lYeWKd3Uat_!!6000000004831-2-tps-1752-600.png)
diff --git a/copaw/tests/test_bridge.py b/copaw/tests/test_bridge.py
index ff6dd451..f2b07983 100644
--- a/copaw/tests/test_bridge.py
+++ b/copaw/tests/test_bridge.py
@@ -14,10 +14,10 @@ def _make_openclaw_cfg(**memory_search_overrides):
         "channels": {"matrix": {"enabled": True, "homeserver": "http://localhost:6167", "accessToken": "tok"}},
         "models": {
             "providers": {
-                "gw": {"baseUrl": "http://aigw:8080/v1", "apiKey": "key123", "models": [{"id": "qwen3.5-plus", "name": "qwen3.5-plus"}]}
+                "gw": {"baseUrl": "http://aigw:8080/v1", "apiKey": "key123", "models": [{"id": "qwen3.6-plus", "name": "qwen3.6-plus"}]}
             }
         },
-        "agents": {"defaults": {"model": {"primary": "gw/qwen3.5-plus"}}},
+        "agents": {"defaults": {"model": {"primary": "gw/qwen3.6-plus"}}},
     }
     if memory_search_overrides is not None:
         base["agents"]["defaults"]["memorySearch"] = {
@@ -61,10 +61,10 @@ def test_bridge_no_embedding_config():
         "channels": {"matrix": {"enabled": True, "homeserver": "http://localhost:6167", "accessToken": "tok"}},
         "models": {
             "providers": {
-                "gw": {"baseUrl": "http://aigw:8080/v1", "apiKey": "key123", "models": [{"id": "qwen3.5-plus", "name": "qwen3.5-plus"}]}
+                "gw": {"baseUrl": "http://aigw:8080/v1", "apiKey": "key123", "models": [{"id": "qwen3.6-plus", "name": "qwen3.6-plus"}]}
             }
         },
-        "agents": {"defaults": {"model": {"primary": "gw/qwen3.5-plus"}}},
+        "agents": {"defaults": {"model": {"primary": "gw/qwen3.6-plus"}}},
     }
     config = _bridge_and_read(openclaw_cfg)
 
diff --git a/docs/declarative-resource-management.md b/docs/declarative-resource-management.md
index 5d5d23c7..c393715a 100644
--- a/docs/declarative-resource-management.md
+++ b/docs/declarative-resource-management.md
@@ -76,7 +76,7 @@ spec:
 | Field | Type | Required | Default | Description |
 |-------|------|----------|---------|-------------|
 | `metadata.name` | string | Yes | — | Worker name, globally unique |
-| `spec.model` | string | Yes | — | LLM model ID, e.g. `claude-sonnet-4-6`, `qwen3.5-plus` |
+| `spec.model` | string | Yes | — | LLM model ID, e.g. `claude-sonnet-4-6`, `qwen3.6-plus` |
 | `spec.runtime` | string | No | `openclaw` | Agent runtime: `openclaw` or `copaw` |
 | `spec.image` | string | No | `hiclaw/worker-agent:latest` | Custom Docker image |
 | `spec.identity` | string | No | — | Worker public identity, used to generate IDENTITY.md |
@@ -734,7 +734,7 @@ kind: Worker
 metadata:
   name: alice
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
   expose:
     - port: 8080
     - port: 3000
@@ -759,14 +759,14 @@ metadata:
 spec:
   leader:
     name: lead
-    model: qwen3.5-plus
+    model: qwen3.6-plus
   workers:
     - name: backend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 8080
     - name: frontend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 3000
 ```
@@ -775,10 +775,10 @@ spec:
 
 ```bash
 # Expose ports via CLI flag
-hiclaw apply worker --name alice --model qwen3.5-plus --expose 8080,3000
+hiclaw apply worker --name alice --model qwen3.6-plus --expose 8080,3000
 
 # Remove exposed ports (re-apply without --expose)
-hiclaw apply worker --name alice --model qwen3.5-plus
+hiclaw apply worker --name alice --model qwen3.6-plus
 ```
 
 ### Use Cases
diff --git a/docs/faq.md b/docs/faq.md
index d2ef0591..1f139eaf 100644
--- a/docs/faq.md
+++ b/docs/faq.md
@@ -126,7 +126,7 @@ HiClaw supports two ways to switch models: **switch the current session model**
 Use the `/model` slash command in IM to instantly switch the model for the current session, no restart needed:
 
 ```
-/model qwen3.5-plus
+/model qwen3.6-plus
 ```
 
 This only affects the current session — the primary model is restored after a restart. Only pre-configured known models are supported; see [`manager/configs/known-models.json`](../manager/configs/known-models.json) for the full list.
@@ -139,7 +139,7 @@ Use Manager's built-in **model-switch skill** to persistently change the primary
 
 **Why use Manager instead of manual config?**
 
-OpenClaw requires setting the model's context window size (`contextWindow`) in its config. HiClaw defaults to qwen3.5-plus's 200K token window. If you switch to a model with a different window without updating this setting, the session may fail when approaching the window limit — OpenClaw won't know when to compress context.
+OpenClaw requires setting the model's context window size (`contextWindow`) in its config. HiClaw defaults to qwen3.6-plus's 1M token window. If you switch to a model with a different window without updating this setting, the session may fail when approaching the window limit — OpenClaw won't know when to compress context.
 
 The model-switch skill:
 1. Looks up the correct `contextWindow` and `maxTokens` for the target model
@@ -176,7 +176,7 @@ Two options are available: **switch the current session model** and **switch the
 In the Worker's group chat or DM, use @mention with the `/model` command to switch instantly:
 
 ```
-@alice /model qwen3.5-plus
+@alice /model qwen3.6-plus
 ```
 
 Only affects the current session — the primary model is restored after a restart. Only pre-configured known models are supported; see [`manager/configs/known-models.json`](../manager/configs/known-models.json) for the full list.
@@ -185,7 +185,7 @@ Only affects the current session — the primary model is restored after a resta
 
 Manager handles this for you, and supports any model name (not limited to the pre-configured list).
 
-**At creation time**: When asking Manager to create a Worker, specify the model name directly, e.g. "Create a Worker named alice using `qwen3.5-plus`."
+**At creation time**: When asking Manager to create a Worker, specify the model name directly, e.g. "Create a Worker named alice using `qwen3.6-plus`."
 
 **After creation**: Tell Manager at any time to switch a Worker's model, e.g. "Switch alice to use `claude-3-5-sonnet`." Manager will update the Worker's configuration accordingly.
 
@@ -197,7 +197,7 @@ Make sure Higress is configured to route the target model name to the correct pr
 
 **Single provider**
 
-In the Higress console, set up `default-ai-route` to route requests to your LLM provider. Then tell Manager the model name you want the Worker to use (e.g. `qwen3.5-plus`). Manager will run a connectivity test with that model name and complete the switch automatically.
+In the Higress console, set up `default-ai-route` to route requests to your LLM provider. Then tell Manager the model name you want the Worker to use (e.g. `qwen3.6-plus`). Manager will run a connectivity test with that model name and complete the switch automatically.
 
 **Multiple providers**
 
diff --git a/docs/import-worker.md b/docs/import-worker.md
index b1bfc67a..bcc5ee9d 100644
--- a/docs/import-worker.md
+++ b/docs/import-worker.md
@@ -377,7 +377,7 @@ bash hiclaw-import.sh -f <resource.yaml> [--prune] [--dry-run]
 | `--name <name>` | Worker name (required) | — |
 | `--zip <path\|url>` | ZIP package (local path or URL) | — |
 | `--package <uri>` | Remote package URI (`nacos://`, `http://`, `oss://`) | — |
-| `--model <model>` | LLM model ID | `qwen3.5-plus` |
+| `--model <model>` | LLM model ID | `qwen3.6-plus` |
 | `--skills <s1,s2>` | Comma-separated built-in skills | — |
 | `--mcp-servers <m1,m2>` | Comma-separated MCP servers | — |
 | `--runtime <runtime>` | Agent runtime (`openclaw`\|`copaw`) | `openclaw` |
diff --git a/docs/manager-guide.md b/docs/manager-guide.md
index 89892415..d42e1bc9 100644
--- a/docs/manager-guide.md
+++ b/docs/manager-guide.md
@@ -16,7 +16,7 @@ The Manager is configured via environment variables set during installation. The
 |----------|----------|---------|-------------|
 | `HICLAW_LLM_API_KEY` | Yes | - | LLM API key |
 | `HICLAW_LLM_PROVIDER` | No | `qwen` | LLM provider (`qwen` for Alibaba Cloud, `openai-compat` for OpenAI-compatible APIs) |
-| `HICLAW_DEFAULT_MODEL` | No | `qwen3.5-plus` | Default model ID |
+| `HICLAW_DEFAULT_MODEL` | No | `qwen3.6-plus` | Default model ID |
 | `HICLAW_ADMIN_USER` | No | `admin` | Human admin Matrix username |
 | `HICLAW_ADMIN_PASSWORD` | No | (auto-generated) | Human admin password (min 8 chars, MinIO requirement) |
 | `HICLAW_MATRIX_DOMAIN` | No | `matrix-local.hiclaw.io:18080` | Matrix server domain (used inside container) |
diff --git a/docs/windows-deploy.md b/docs/windows-deploy.md
index 258e39c3..d93b93a8 100644
--- a/docs/windows-deploy.md
+++ b/docs/windows-deploy.md
@@ -194,7 +194,7 @@ If you selected CodingPlan in the previous step, a model selection menu appears:
 
 ```
 Select CodingPlan Default Model:
-  1) qwen3.5-plus  - Qwen 3.5 (Fastest)
+  1) qwen3.6-plus  - Qwen 3.6 (Fastest)
   2) glm-5  - Zhipu GLM-5 (Recommended for coding)
   3) kimi-k2.5  - Moonshot Kimi K2.5
   4) MiniMax-M2.5  - MiniMax M2.5
diff --git a/docs/zh-cn/declarative-resource-management.md b/docs/zh-cn/declarative-resource-management.md
index d36642c6..6e39e0bc 100644
--- a/docs/zh-cn/declarative-resource-management.md
+++ b/docs/zh-cn/declarative-resource-management.md
@@ -76,7 +76,7 @@ spec:
 | 字段 | 类型 | 必填 | 默认值 | 说明 |
 |------|------|------|--------|------|
 | `metadata.name` | string | 是 | — | Worker 名称，全局唯一 |
-| `spec.model` | string | 是 | — | LLM 模型 ID，如 `claude-sonnet-4-6`、`qwen3.5-plus` |
+| `spec.model` | string | 是 | — | LLM 模型 ID，如 `claude-sonnet-4-6`、`qwen3.6-plus` |
 | `spec.runtime` | string | 否 | `openclaw` | Agent 运行时，`openclaw` 或 `copaw` |
 | `spec.image` | string | 否 | `hiclaw/worker-agent:latest` | 自定义 Docker 镜像 |
 | `spec.identity` | string | 否 | — | Worker 公开身份信息，用于生成 IDENTITY.md |
@@ -734,7 +734,7 @@ kind: Worker
 metadata:
   name: alice
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
   expose:
     - port: 8080
     - port: 3000
@@ -759,14 +759,14 @@ metadata:
 spec:
   leader:
     name: lead
-    model: qwen3.5-plus
+    model: qwen3.6-plus
   workers:
     - name: backend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 8080
     - name: frontend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 3000
 ```
@@ -775,10 +775,10 @@ spec:
 
 ```bash
 # 通过 CLI 参数暴露端口
-hiclaw apply worker --name alice --model qwen3.5-plus --expose 8080,3000
+hiclaw apply worker --name alice --model qwen3.6-plus --expose 8080,3000
 
 # 取消暴露（不带 --expose 重新 apply）
-hiclaw apply worker --name alice --model qwen3.5-plus
+hiclaw apply worker --name alice --model qwen3.6-plus
 ```
 
 ### 使用场景
diff --git a/docs/zh-cn/faq.md b/docs/zh-cn/faq.md
index 2e55588c..9c78b1c4 100644
--- a/docs/zh-cn/faq.md
+++ b/docs/zh-cn/faq.md
@@ -123,7 +123,7 @@ HiClaw 支持两种模型切换方式：**切换当前会话模型**（即时生
 通过 IM 斜杠命令 `/model` 可以即时切换当前会话使用的模型，无需重启：
 
 ```
-/model qwen3.5-plus
+/model qwen3.6-plus
 ```
 
 这种方式仅对当前会话生效，重启后会恢复为主用模型。且仅支持预置的已知模型列表中的模型，完整列表见 [`manager/configs/known-models.json`](../../manager/configs/known-models.json)。
@@ -136,7 +136,7 @@ HiClaw 支持两种模型切换方式：**切换当前会话模型**（即时生
 
 **为什么让 Manager 切换而不是手动改配置？**
 
-OpenClaw 需要在配置中设置模型的上下文窗口大小（`contextWindow`）。HiClaw 默认使用 qwen3.5-plus 的 200K token 窗口。如果切换到窗口不同的模型但没有更新这个设置，当对话接近窗口上限时，OpenClaw 不知道何时压缩上下文，可能导致 session 无法使用。
+OpenClaw 需要在配置中设置模型的上下文窗口大小（`contextWindow`）。HiClaw 默认使用 qwen3.6-plus 的 1M token 窗口。如果切换到窗口不同的模型但没有更新这个设置，当对话接近窗口上限时，OpenClaw 不知道何时压缩上下文，可能导致 session 无法使用。
 
 模型切换技能会根据模型名自动修改 OpenClaw 配置中的 `contextWindow` 和 `maxTokens`。
 
@@ -158,7 +158,7 @@ Manager 会使用模型切换技能完成配置更新。
 
 **单供应商情况**
 
-在 Higress 控制台，将 `default-ai-route` 这个路由配置到你的模型供应商。然后直接告诉 Manager 你想使用的具体模型名（例如 `qwen3.5-plus`）。Manager 会先用该模型名发起一次联通测试，测试通过后自动完成切换。
+在 Higress 控制台，将 `default-ai-route` 这个路由配置到你的模型供应商。然后直接告诉 Manager 你想使用的具体模型名（例如 `qwen3.6-plus`）。Manager 会先用该模型名发起一次联通测试，测试通过后自动完成切换。
 
 **多供应商情况**
 
@@ -177,7 +177,7 @@ Manager 会使用模型切换技能完成配置更新。
 在 Worker 所在的群聊或私聊中，通过 @Worker 加 `/model` 命令即时切换：
 
 ```
-@alice /model qwen3.5-plus
+@alice /model qwen3.6-plus
 ```
 
 仅对当前会话生效，重启后恢复为主用模型。仅支持预置的已知模型，完整列表见 [`manager/configs/known-models.json`](../../manager/configs/known-models.json)。
@@ -186,7 +186,7 @@ Manager 会使用模型切换技能完成配置更新。
 
 由 Manager 代为操作，支持任意模型名。
 
-**创建时指定**：在让 Manager 创建 Worker 时直接说明模型，例如"帮我创建一个名为 alice 的 Worker，使用 `qwen3.5-plus`"。
+**创建时指定**：在让 Manager 创建 Worker 时直接说明模型，例如"帮我创建一个名为 alice 的 Worker，使用 `qwen3.6-plus`"。
 
 **创建后修改**：随时告诉 Manager 切换某个 Worker 的模型，例如"把 alice 的模型切换为 `claude-3-5-sonnet`"，Manager 会自动更新该 Worker 的配置。
 
@@ -198,7 +198,7 @@ Manager 会使用模型切换技能完成配置更新。
 
 **单供应商情况**
 
-在 Higress 控制台，将 `default-ai-route` 这个路由配置到你的模型供应商。然后直接告诉 Manager 你想让 Worker 使用的具体模型名（例如 `qwen3.5-plus`）。Manager 会先用该模型名发起一次联通测试，测试通过后自动完成切换。
+在 Higress 控制台，将 `default-ai-route` 这个路由配置到你的模型供应商。然后直接告诉 Manager 你想让 Worker 使用的具体模型名（例如 `qwen3.6-plus`）。Manager 会先用该模型名发起一次联通测试，测试通过后自动完成切换。
 
 **多供应商情况**
 
diff --git a/docs/zh-cn/import-worker.md b/docs/zh-cn/import-worker.md
index 9cbaf315..3654fbd4 100644
--- a/docs/zh-cn/import-worker.md
+++ b/docs/zh-cn/import-worker.md
@@ -375,7 +375,7 @@ bash hiclaw-import.sh -f <resource.yaml> [--prune] [--dry-run]
 | `--name <名称>` | Worker 名称（必需） | — |
 | `--zip <路径\|URL>` | ZIP 包（本地路径或 URL） | — |
 | `--package <URI>` | 远程包 URI（`nacos://`、`http://`、`oss://`） | — |
-| `--model <模型>` | LLM 模型 ID | `qwen3.5-plus` |
+| `--model <模型>` | LLM 模型 ID | `qwen3.6-plus` |
 | `--skills <s1,s2>` | 逗号分隔的内置技能 | — |
 | `--mcp-servers <m1,m2>` | 逗号分隔的 MCP Server | — |
 | `--runtime <运行时>` | Agent 运行时（`openclaw`\|`copaw`） | `openclaw` |
diff --git a/docs/zh-cn/manager-guide.md b/docs/zh-cn/manager-guide.md
index c8d239e2..13d800f6 100644
--- a/docs/zh-cn/manager-guide.md
+++ b/docs/zh-cn/manager-guide.md
@@ -16,7 +16,7 @@ Manager 通过安装时设置的环境变量进行配置。安装脚本会生成
 |------|----------|--------|------|
 | `HICLAW_LLM_API_KEY` | 是 | - | LLM API Key |
 | `HICLAW_LLM_PROVIDER` | 否 | `qwen` | LLM 提供商（`qwen` 为阿里云百炼，`openai-compat` 为 OpenAI 兼容 API） |
-| `HICLAW_DEFAULT_MODEL` | 否 | `qwen3.5-plus` | 默认模型 ID |
+| `HICLAW_DEFAULT_MODEL` | 否 | `qwen3.6-plus` | 默认模型 ID |
 | `HICLAW_ADMIN_USER` | 否 | `admin` | 人工管理员的 Matrix 用户名 |
 | `HICLAW_ADMIN_PASSWORD` | 否 | （自动生成） | 管理员密码（最少 8 位，MinIO 要求） |
 | `HICLAW_MATRIX_DOMAIN` | 否 | `matrix-local.hiclaw.io:18080` | Matrix 服务器域名（容器内使用） |
diff --git a/docs/zh-cn/windows-deploy.md b/docs/zh-cn/windows-deploy.md
index 10becc12..c2153791 100644
--- a/docs/zh-cn/windows-deploy.md
+++ b/docs/zh-cn/windows-deploy.md
@@ -198,7 +198,7 @@ Set-ExecutionPolicy Bypass -Scope Process -Force; $wc=New-Object Net.WebClient;
 
 ```
 选择 CodingPlan 默认模型:
-  1) qwen3.5-plus  - 千问 3.5（速度最快）
+  1) qwen3.6-plus  - 千问 3.6（速度最快）
   2) glm-5  - 智谱 GLM-5（编程推荐）
   3) kimi-k2.5  - Moonshot Kimi K2.5
   4) MiniMax-M2.5  - MiniMax M2.5
diff --git a/hiclaw-controller/cmd/hiclaw/main.go b/hiclaw-controller/cmd/hiclaw/main.go
index 3bc8789a..b6d8dee0 100644
--- a/hiclaw-controller/cmd/hiclaw/main.go
+++ b/hiclaw-controller/cmd/hiclaw/main.go
@@ -140,9 +140,9 @@ func applyWorkerCmd() *cobra.Command {
   hiclaw apply worker --name alice --model claude-sonnet-4-6 --package nacos://inst/ns/spec/v1
   hiclaw apply worker --name alice --package reviewer
   hiclaw apply worker --name alice --package reviewer/label:latest
-  hiclaw apply worker --name bob --model qwen3.5-plus
+  hiclaw apply worker --name bob --model qwen3.6-plus
   hiclaw apply worker --name charlie --model gpt-5-mini --skills github-operations --mcp-servers github
-  hiclaw apply worker --name alice --model qwen3.5-plus --expose 8080,3000`,
+  hiclaw apply worker --name alice --model qwen3.6-plus --expose 8080,3000`,
 		RunE: func(cmd *cobra.Command, args []string) error {
 			if name == "" {
 				return fmt.Errorf("--name is required")
@@ -155,7 +155,7 @@ func applyWorkerCmd() *cobra.Command {
 
 			// Generate Worker YAML from CLI params
 			if model == "" {
-				model = "qwen3.5-plus"
+				model = "qwen3.6-plus"
 			}
 
 			return applyWorkerFromParams(name, model, packageURI, skills, mcpServers, runtime, expose, dryRun)
@@ -163,7 +163,7 @@ func applyWorkerCmd() *cobra.Command {
 	}
 
 	cmd.Flags().StringVar(&name, "name", "", "Worker name (required)")
-	cmd.Flags().StringVar(&model, "model", "", "LLM model ID (default: qwen3.5-plus)")
+	cmd.Flags().StringVar(&model, "model", "", "LLM model ID (default: qwen3.6-plus)")
 	cmd.Flags().StringVar(&zipFile, "zip", "", "Local ZIP package (manifest.json)")
 	cmd.Flags().StringVar(&packageURI, "package", "", "Remote package URI (nacos://, http://, oss://) or Nacos shorthand (name, name/version, name/label:latest)")
 	cmd.Flags().StringVar(&skills, "skills", "", "Comma-separated built-in skills")
@@ -712,7 +712,7 @@ func applyZip(zipPath string, name string, dryRun bool) error {
 
 	model := jsonField(string(manifestData), "model")
 	if model == "" {
-		model = "qwen3.5-plus"
+		model = "qwen3.6-plus"
 	}
 
 	// Compute SHA256 of ZIP for content-addressable storage
diff --git a/hiclaw-controller/cmd/hiclaw/main_test.go b/hiclaw-controller/cmd/hiclaw/main_test.go
index bb5022e5..03c1c176 100644
--- a/hiclaw-controller/cmd/hiclaw/main_test.go
+++ b/hiclaw-controller/cmd/hiclaw/main_test.go
@@ -167,7 +167,7 @@ kind: Worker
 metadata:
   name: bob
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
 `
 	tmpFile := writeTempYAMLForTest(t, yaml)
 	resources, err := loadResources([]string{tmpFile})
@@ -366,7 +366,7 @@ metadata:
   name: bob
 spec:
   package: https://example.com/worker.zip
-  model: qwen3.5-plus
+  model: qwen3.6-plus
 `,
 			expected: "https://example.com/worker.zip",
 		},
@@ -377,7 +377,7 @@ kind: Worker
 metadata:
   name: charlie
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
 `,
 			expected: "",
 		},
diff --git a/hiclaw-controller/config/crd/workers.hiclaw.io.yaml b/hiclaw-controller/config/crd/workers.hiclaw.io.yaml
index 753f9813..37bcd431 100644
--- a/hiclaw-controller/config/crd/workers.hiclaw.io.yaml
+++ b/hiclaw-controller/config/crd/workers.hiclaw.io.yaml
@@ -18,7 +18,7 @@ spec:
               properties:
                 model:
                   type: string
-                  description: LLM model ID (e.g. claude-sonnet-4-6, qwen3.5-plus)
+                  description: LLM model ID (e.g. claude-sonnet-4-6, qwen3.6-plus)
                 runtime:
                   type: string
                   enum: [openclaw, copaw]
diff --git a/install/README.md b/install/README.md
index eb795118..b24f0877 100644
--- a/install/README.md
+++ b/install/README.md
@@ -138,7 +138,7 @@ $env:HICLAW_LLM_API_KEY = "your-api-key"
 |----------|-------------|---------|
 | `HICLAW_NON_INTERACTIVE` | Skip all prompts | `0` |
 | `HICLAW_LLM_PROVIDER` | LLM provider (`qwen` or `openai-compat`) | `qwen` |
-| `HICLAW_DEFAULT_MODEL` | Default model ID | `qwen3.5-plus` |
+| `HICLAW_DEFAULT_MODEL` | Default model ID | `qwen3.6-plus` |
 | `HICLAW_LLM_API_KEY` | LLM API key | *(required)* |
 | `HICLAW_ADMIN_USER` | Admin username | `admin` |
 | `HICLAW_ADMIN_PASSWORD` | Admin password | *(auto-generated)* |
diff --git a/install/hiclaw-install.ps1 b/install/hiclaw-install.ps1
index e2282b9b..a929b0b8 100644
--- a/install/hiclaw-install.ps1
+++ b/install/hiclaw-install.ps1
@@ -14,7 +14,7 @@
 # Environment variables (for automation):
 #   HICLAW_NON_INTERACTIVE    Skip all prompts, use defaults  (default: 0)
 #   HICLAW_LLM_PROVIDER       LLM provider       (default: qwen)
-#   HICLAW_DEFAULT_MODEL      Default model      (default: qwen3.5-plus)
+#   HICLAW_DEFAULT_MODEL      Default model      (default: qwen3.6-plus)
 #   HICLAW_LLM_API_KEY        LLM API key        (required)
 #   HICLAW_ADMIN_USER         Admin username     (default: admin)
 #   HICLAW_ADMIN_PASSWORD     Admin password     (auto-generated if not set, min 8 chars)
@@ -297,7 +297,7 @@ $script:Messages = @{
     "llm.alibaba.model.qwen" = @{ zh = "  2) 百炼通用接口"; en = "  2) qwen general  - General purpose LLM" }
     "llm.alibaba.model.select" = @{ zh = "选择模型系列 [1/2]"; en = "Select model series [1/2]" }
     "llm.codingplan.models_title" = @{ zh = "选择 CodingPlan 默认模型:"; en = "Select CodingPlan default model:" }
-    "llm.codingplan.model.qwen35plus" = @{ zh = "  1) qwen3.5-plus  - 千问 3.5（速度最快）"; en = "  1) qwen3.5-plus  - Qwen 3.5 (fastest)" }
+    "llm.codingplan.model.qwen36plus" = @{ zh = "  1) qwen3.6-plus  - 千问 3.6（速度最快）"; en = "  1) qwen3.6-plus  - Qwen 3.6 (fastest)" }
     "llm.codingplan.model.glm5" = @{ zh = "  2) glm-5  - 智谱 GLM-5（编程推荐）"; en = "  2) glm-5  - Zhipu GLM-5 (recommended for coding)" }
     "llm.codingplan.model.kimi" = @{ zh = "  3) kimi-k2.5  - Moonshot Kimi K2.5"; en = "  3) kimi-k2.5  - Moonshot Kimi K2.5" }
     "llm.codingplan.model.minimax" = @{ zh = "  4) MiniMax-M2.5  - MiniMax M2.5"; en = "  4) MiniMax-M2.5  - MiniMax M2.5" }
@@ -306,7 +306,7 @@ $script:Messages = @{
     "llm.provider.selected_qwen" = @{ zh = "  提供商: 阿里云百炼"; en = "  Provider: Alibaba Cloud Bailian" }
     "llm.provider.selected_openai" = @{ zh = "  提供商: {0}（OpenAI 兼容）"; en = "  Provider: {0} (OpenAI-compatible)" }
     "llm.provider.invalid" = @{ zh = "无效选择: {0}（请输入 1 或 2）"; en = "Invalid choice: {0} (please enter 1 or 2)" }
-    "llm.qwen.model_prompt" = @{ zh = "默认模型 ID [qwen3.5-plus]"; en = "Default Model ID [qwen3.5-plus]" }
+    "llm.qwen.model_prompt" = @{ zh = "默认模型 ID [qwen3.6-plus]"; en = "Default Model ID [qwen3.6-plus]" }
     "llm.openai.base_url_prompt" = @{ zh = "Base URL（例如 https://api.openai.com/v1）"; en = "Base URL (e.g., https://api.openai.com/v1)" }
     "llm.openai.model_prompt" = @{ zh = "默认模型 ID [gpt-5.4]"; en = "Default Model ID [gpt-5.4]" }
     "llm.openai.base_url_label" = @{ zh = "  Base URL: {0}"; en = "  Base URL: {0}" }
@@ -639,7 +639,7 @@ function Get-LanIP {
 $script:KnownModels = @(
     "gpt-5.4", "gpt-5.3-codex", "gpt-5-mini", "gpt-5-nano",
     "claude-opus-4-6", "claude-sonnet-4-6", "claude-haiku-4-5",
-    "qwen3.5-plus", "deepseek-chat", "deepseek-reasoner",
+    "qwen3.5-plus", "qwen3.6-plus", "deepseek-chat", "deepseek-reasoner",
     "kimi-k2.5", "glm-5", "MiniMax-M2.7", "MiniMax-M2.7-highspeed", "MiniMax-M2.5"
 )
 
@@ -1401,7 +1401,7 @@ function Step-Llm {
 
     if ($script:HICLAW_NON_INTERACTIVE) {
         $script:config.LLM_PROVIDER = if ($env:HICLAW_LLM_PROVIDER) { $env:HICLAW_LLM_PROVIDER } else { "qwen" }
-        $script:config.DEFAULT_MODEL = if ($env:HICLAW_DEFAULT_MODEL) { $env:HICLAW_DEFAULT_MODEL } else { "qwen3.5-plus" }
+        $script:config.DEFAULT_MODEL = if ($env:HICLAW_DEFAULT_MODEL) { $env:HICLAW_DEFAULT_MODEL } else { "qwen3.6-plus" }
         $script:config.OPENAI_BASE_URL = if ($env:HICLAW_OPENAI_BASE_URL) { $env:HICLAW_OPENAI_BASE_URL } else { "" }
         Write-Log (Get-Msg "llm.provider.label" -f $script:config.LLM_PROVIDER)
         Write-Log (Get-Msg "llm.model.label" -f $script:config.DEFAULT_MODEL)
@@ -1433,7 +1433,7 @@ function Step-Llm {
                 $modelChoice = "codingplan"
                 Write-Host ""
                 Write-Host (Get-Msg "llm.codingplan.models_title")
-                Write-Host (Get-Msg "llm.codingplan.model.qwen35plus")
+                Write-Host (Get-Msg "llm.codingplan.model.qwen36plus")
                 Write-Host (Get-Msg "llm.codingplan.model.glm5")
                 Write-Host (Get-Msg "llm.codingplan.model.kimi")
                 Write-Host (Get-Msg "llm.codingplan.model.minimax")
@@ -1446,11 +1446,11 @@ function Step-Llm {
                 $codingPlanModelChoice = if ($codingPlanModelChoice) { $codingPlanModelChoice } else { "1" }
                 if ($codingPlanModelChoice -eq "b") { $script:StepResult = "back"; return }
                 switch -Regex ($codingPlanModelChoice) {
-                    "^(1|qwen3\.5-plus)$"  { $script:config.DEFAULT_MODEL = "qwen3.5-plus" }
+                    "^(1|qwen3\.6-plus)$"  { $script:config.DEFAULT_MODEL = "qwen3.6-plus" }
                     "^(2|glm-5)$"          { $script:config.DEFAULT_MODEL = "glm-5" }
                     "^(3|kimi-k2\.5)$"     { $script:config.DEFAULT_MODEL = "kimi-k2.5" }
                     "^(4|MiniMax-M2\.5)$"  { $script:config.DEFAULT_MODEL = "MiniMax-M2.5" }
-                    default                { $script:config.DEFAULT_MODEL = "qwen3.5-plus" }
+                    default                { $script:config.DEFAULT_MODEL = "qwen3.6-plus" }
                 }
                 Write-Log (Get-Msg "llm.provider.selected_codingplan")
             } else {
@@ -1472,7 +1472,7 @@ function Step-Llm {
                     Write-Host ""
                     $qwenModelInput = Read-Host (Get-Msg "llm.qwen.model_prompt")
                     if ($qwenModelInput -eq "b") { $script:StepResult = "back"; return }
-                    $script:config.DEFAULT_MODEL = if ($qwenModelInput) { $qwenModelInput } elseif ($env:HICLAW_DEFAULT_MODEL) { $env:HICLAW_DEFAULT_MODEL } else { "qwen3.5-plus" }
+                    $script:config.DEFAULT_MODEL = if ($qwenModelInput) { $qwenModelInput } elseif ($env:HICLAW_DEFAULT_MODEL) { $env:HICLAW_DEFAULT_MODEL } else { "qwen3.6-plus" }
                     $script:config.OPENAI_BASE_URL = ""
                     Write-Log (Get-Msg "llm.provider.selected_qwen")
                     Request-CustomModelParams $script:config.DEFAULT_MODEL
@@ -1482,7 +1482,7 @@ function Step-Llm {
                     $script:config.OPENAI_BASE_URL = "https://coding.dashscope.aliyuncs.com/v1"
                     Write-Host ""
                     Write-Host (Get-Msg "llm.codingplan.models_title")
-                    Write-Host (Get-Msg "llm.codingplan.model.qwen35plus")
+                    Write-Host (Get-Msg "llm.codingplan.model.qwen36plus")
                     Write-Host (Get-Msg "llm.codingplan.model.glm5")
                     Write-Host (Get-Msg "llm.codingplan.model.kimi")
                     Write-Host (Get-Msg "llm.codingplan.model.minimax")
@@ -1495,11 +1495,11 @@ function Step-Llm {
                     $codingPlanModelChoice = if ($codingPlanModelChoice) { $codingPlanModelChoice } else { "1" }
                     if ($codingPlanModelChoice -eq "b") { $script:StepResult = "back"; return }
                     switch -Regex ($codingPlanModelChoice) {
-                        "^(1|qwen3\.5-plus)$"  { $script:config.DEFAULT_MODEL = "qwen3.5-plus" }
+                        "^(1|qwen3\.6-plus)$"  { $script:config.DEFAULT_MODEL = "qwen3.6-plus" }
                         "^(2|glm-5)$"          { $script:config.DEFAULT_MODEL = "glm-5" }
                         "^(3|kimi-k2\.5)$"     { $script:config.DEFAULT_MODEL = "kimi-k2.5" }
                         "^(4|MiniMax-M2\.5)$"  { $script:config.DEFAULT_MODEL = "MiniMax-M2.5" }
-                        default                { $script:config.DEFAULT_MODEL = "qwen3.5-plus" }
+                        default                { $script:config.DEFAULT_MODEL = "qwen3.6-plus" }
                     }
                     Write-Log (Get-Msg "llm.provider.selected_codingplan")
                 }
diff --git a/install/hiclaw-install.sh b/install/hiclaw-install.sh
index 9f810c2b..09fd287a 100644
--- a/install/hiclaw-install.sh
+++ b/install/hiclaw-install.sh
@@ -13,7 +13,7 @@
 # Environment variables (for automation):
 #   HICLAW_NON_INTERACTIVE    Skip all prompts, use defaults  (default: 0)
 #   HICLAW_LLM_PROVIDER      LLM provider       (default: alibaba-cloud)
-#   HICLAW_DEFAULT_MODEL      Default model       (default: qwen3.5-plus)
+#   HICLAW_DEFAULT_MODEL      Default model       (default: qwen3.6-plus)
 #   HICLAW_LLM_API_KEY        LLM API key         (required)
 #   HICLAW_ADMIN_USER         Admin username       (default: admin)
 #   HICLAW_ADMIN_PASSWORD     Admin password       (auto-generated if not set, min 8 chars)
@@ -381,8 +381,8 @@ msg() {
         "llm.alibaba.model.select.en") text="Select model series [1/2]" ;;
         "llm.codingplan.models_title.zh") text="选择 CodingPlan 默认模型:" ;;
         "llm.codingplan.models_title.en") text="Select CodingPlan default model:" ;;
-        "llm.codingplan.model.qwen35plus.zh") text="  1) qwen3.5-plus  - 千问 3.5（速度最快）" ;;
-        "llm.codingplan.model.qwen35plus.en") text="  1) qwen3.5-plus  - Qwen 3.5 (fastest)" ;;
+        "llm.codingplan.model.qwen36plus.zh") text="  1) qwen3.6-plus  - 千问 3.6（速度最快）" ;;
+        "llm.codingplan.model.qwen36plus.en") text="  1) qwen3.6-plus  - Qwen 3.6 (fastest)" ;;
         "llm.codingplan.model.glm5.zh") text="  2) glm-5  - 智谱 GLM-5（编程推荐）" ;;
         "llm.codingplan.model.glm5.en") text="  2) glm-5  - Zhipu GLM-5 (recommended for coding)" ;;
         "llm.codingplan.model.kimi.zh") text="  3) kimi-k2.5  - Moonshot Kimi K2.5" ;;
@@ -399,8 +399,8 @@ msg() {
         "llm.provider.selected_openai.en") text="  Provider: %s (OpenAI-compatible)" ;;
         "llm.provider.invalid.zh") text="无效选择: %s（请输入 1 或 2）" ;;
         "llm.provider.invalid.en") text="Invalid choice: %s (please enter 1 or 2)" ;;
-        "llm.qwen.model_prompt.zh") text="默认模型 ID [qwen3.5-plus]" ;;
-        "llm.qwen.model_prompt.en") text="Default Model ID [qwen3.5-plus]" ;;
+        "llm.qwen.model_prompt.zh") text="默认模型 ID [qwen3.6-plus]" ;;
+        "llm.qwen.model_prompt.en") text="Default Model ID [qwen3.6-plus]" ;;
         "llm.openai.base_url_prompt.zh") text="Base URL（例如 https://api.openai.com/v1）" ;;
         "llm.openai.base_url_prompt.en") text="Base URL (e.g., https://api.openai.com/v1)" ;;
         "llm.openai.model_prompt.zh") text="默认模型 ID [gpt-5.4]" ;;
@@ -926,7 +926,7 @@ resolve_docker_proxy_image() {
 # ============================================================
 # Known models list — used to detect custom models during install
 # ============================================================
-KNOWN_MODELS="gpt-5.4 gpt-5.3-codex gpt-5-mini gpt-5-nano claude-opus-4-6 claude-sonnet-4-6 claude-haiku-4-5 qwen3.5-plus deepseek-chat deepseek-reasoner kimi-k2.5 glm-5 MiniMax-M2.7 MiniMax-M2.7-highspeed MiniMax-M2.5"
+KNOWN_MODELS="gpt-5.4 gpt-5.3-codex gpt-5-mini gpt-5-nano claude-opus-4-6 claude-sonnet-4-6 claude-haiku-4-5 qwen3.5-plus qwen3.6-plus deepseek-chat deepseek-reasoner kimi-k2.5 glm-5 MiniMax-M2.7 MiniMax-M2.7-highspeed MiniMax-M2.5"
 
 is_known_model() {
     local model="$1"
@@ -1556,7 +1556,7 @@ step_llm() {
     log "$(msg llm.title)"
     if [ "${HICLAW_NON_INTERACTIVE}" = "1" ]; then
         HICLAW_LLM_PROVIDER="${HICLAW_LLM_PROVIDER:-qwen}"
-        HICLAW_DEFAULT_MODEL="${HICLAW_DEFAULT_MODEL:-qwen3.5-plus}"
+        HICLAW_DEFAULT_MODEL="${HICLAW_DEFAULT_MODEL:-qwen3.6-plus}"
         log "$(msg llm.provider.qwen_default "${HICLAW_LLM_PROVIDER}")"
         log "$(msg llm.model.default "${HICLAW_DEFAULT_MODEL}")"
         prompt HICLAW_LLM_API_KEY "$(msg llm.apikey_prompt)" "" "true"
@@ -1586,7 +1586,7 @@ step_llm() {
                 ALIBABA_MODEL_CHOICE="codingplan"
                 echo ""
                 echo "$(msg llm.codingplan.models_title)"
-                echo "$(msg llm.codingplan.model.qwen35plus)"
+                echo "$(msg llm.codingplan.model.qwen36plus)"
                 echo "$(msg llm.codingplan.model.glm5)"
                 echo "$(msg llm.codingplan.model.kimi)"
                 echo "$(msg llm.codingplan.model.minimax)"
@@ -1601,11 +1601,11 @@ step_llm() {
                 fi
                 if [ "${CODINGPLAN_MODEL_CHOICE}" = "b" ]; then STEP_RESULT="back"; return 0; fi
                 case "${CODINGPLAN_MODEL_CHOICE}" in
-                    1|qwen3.5-plus) HICLAW_DEFAULT_MODEL="qwen3.5-plus" ;;
+                    1|qwen3.6-plus) HICLAW_DEFAULT_MODEL="qwen3.6-plus" ;;
                     2|glm-5)        HICLAW_DEFAULT_MODEL="glm-5" ;;
                     3|kimi-k2.5)    HICLAW_DEFAULT_MODEL="kimi-k2.5" ;;
                     4|MiniMax-M2.5) HICLAW_DEFAULT_MODEL="MiniMax-M2.5" ;;
-                    *)              HICLAW_DEFAULT_MODEL="qwen3.5-plus" ;;
+                    *)              HICLAW_DEFAULT_MODEL="qwen3.6-plus" ;;
                 esac
                 log "$(msg llm.provider.selected_codingplan)"
                 log "$(msg llm.model.label "${HICLAW_DEFAULT_MODEL}")"
@@ -1630,7 +1630,7 @@ step_llm() {
                         echo ""
                         read -e -p "$(msg llm.qwen.model_prompt): " HICLAW_DEFAULT_MODEL
                         if [ "${HICLAW_DEFAULT_MODEL}" = "b" ]; then STEP_RESULT="back"; return 0; fi
-                        HICLAW_DEFAULT_MODEL="${HICLAW_DEFAULT_MODEL:-qwen3.5-plus}"
+                        HICLAW_DEFAULT_MODEL="${HICLAW_DEFAULT_MODEL:-qwen3.6-plus}"
                         log "$(msg llm.provider.selected_qwen)"
                         log "$(msg llm.model.label "${HICLAW_DEFAULT_MODEL}")"
                         prompt_custom_model_params "${HICLAW_DEFAULT_MODEL}" || return 0
@@ -1640,7 +1640,7 @@ step_llm() {
                         HICLAW_OPENAI_BASE_URL="https://coding.dashscope.aliyuncs.com/v1"
                         echo ""
                         echo "$(msg llm.codingplan.models_title)"
-                        echo "$(msg llm.codingplan.model.qwen35plus)"
+                        echo "$(msg llm.codingplan.model.qwen36plus)"
                         echo "$(msg llm.codingplan.model.glm5)"
                         echo "$(msg llm.codingplan.model.kimi)"
                         echo "$(msg llm.codingplan.model.minimax)"
@@ -1655,11 +1655,11 @@ step_llm() {
                         fi
                         if [ "${CODINGPLAN_MODEL_CHOICE}" = "b" ]; then STEP_RESULT="back"; return 0; fi
                         case "${CODINGPLAN_MODEL_CHOICE}" in
-                            1|qwen3.5-plus) HICLAW_DEFAULT_MODEL="qwen3.5-plus" ;;
+                            1|qwen3.6-plus) HICLAW_DEFAULT_MODEL="qwen3.6-plus" ;;
                             2|glm-5)        HICLAW_DEFAULT_MODEL="glm-5" ;;
                             3|kimi-k2.5)    HICLAW_DEFAULT_MODEL="kimi-k2.5" ;;
                             4|MiniMax-M2.5) HICLAW_DEFAULT_MODEL="MiniMax-M2.5" ;;
-                            *)              HICLAW_DEFAULT_MODEL="qwen3.5-plus" ;;
+                            *)              HICLAW_DEFAULT_MODEL="qwen3.6-plus" ;;
                         esac
                         log "$(msg llm.provider.selected_codingplan)"
                         log "$(msg llm.model.label "${HICLAW_DEFAULT_MODEL}")"
diff --git a/manager/agent/skills/model-switch/SKILL.md b/manager/agent/skills/model-switch/SKILL.md
index 6560976a..622abcdf 100644
--- a/manager/agent/skills/model-switch/SKILL.md
+++ b/manager/agent/skills/model-switch/SKILL.md
@@ -78,5 +78,6 @@ When the human admin requests switching to a model you don't recognize, you MUST
 | claude-sonnet-4-6 | 1,000,000 | 64,000 |
 | claude-haiku-4-5 | 200,000 | 64,000 |
 | qwen3.5-plus | 200,000 | 64,000 |
+| qwen3.6-plus | 1,000,000 | 64,000 |
 | deepseek-chat / deepseek-reasoner / kimi-k2.5 | 256,000 | 128,000 |
 | glm-5 / MiniMax-M2.7 / MiniMax-M2.7-highspeed / MiniMax-M2.5 | 200,000 | 128,000 |
diff --git a/manager/agent/skills/model-switch/scripts/update-manager-model.sh b/manager/agent/skills/model-switch/scripts/update-manager-model.sh
index 4b799994..4bb219f2 100755
--- a/manager/agent/skills/model-switch/scripts/update-manager-model.sh
+++ b/manager/agent/skills/model-switch/scripts/update-manager-model.sh
@@ -97,6 +97,8 @@ case "${MODEL_NAME}" in
         CTX=200000; MAX=64000 ;;
     qwen3.5-plus)
         CTX=200000; MAX=64000 ;;
+    qwen3.6-plus)
+        CTX=1000000; MAX=64000 ;;
     deepseek-chat|deepseek-reasoner|kimi-k2.5)
         CTX=256000; MAX=128000 ;;
     glm-5|MiniMax-M2.7|MiniMax-M2.7-highspeed|MiniMax-M2.5)
@@ -112,7 +114,7 @@ fi
 
 # Resolve input modalities: only vision-capable models get "image"
 case "${MODEL_NAME}" in
-    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|kimi-k2.5)
+    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|qwen3.6-plus|kimi-k2.5)
         INPUT='["text", "image"]' ;;
     *)
         INPUT='["text"]' ;;
diff --git a/manager/agent/skills/service-publishing/SKILL.md b/manager/agent/skills/service-publishing/SKILL.md
index ea33c6cc..0d3e68a8 100644
--- a/manager/agent/skills/service-publishing/SKILL.md
+++ b/manager/agent/skills/service-publishing/SKILL.md
@@ -26,17 +26,17 @@ Example: worker `alice` exposing port `8080` → `worker-alice-8080-local.hiclaw
 
 ```bash
 # Expose port 8080 for worker alice
-hiclaw apply worker --name alice --model qwen3.5-plus --expose 8080
+hiclaw apply worker --name alice --model qwen3.6-plus --expose 8080
 
 # Expose multiple ports
-hiclaw apply worker --name alice --model qwen3.5-plus --expose 8080,3000
+hiclaw apply worker --name alice --model qwen3.6-plus --expose 8080,3000
 
 # Check exposed ports
 hiclaw get worker alice
 # Look for status.exposedPorts in the output
 
 # Remove exposed ports (update without --expose)
-hiclaw apply worker --name alice --model qwen3.5-plus
+hiclaw apply worker --name alice --model qwen3.6-plus
 ```
 
 ### Via YAML
@@ -47,7 +47,7 @@ kind: Worker
 metadata:
   name: alice
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
   expose:
     - port: 8080
     - port: 3000
@@ -70,14 +70,14 @@ metadata:
 spec:
   leader:
     name: lead
-    model: qwen3.5-plus
+    model: qwen3.6-plus
   workers:
     - name: backend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 8080
     - name: frontend
-      model: qwen3.5-plus
+      model: qwen3.6-plus
       expose:
         - port: 3000
 ```
diff --git a/manager/agent/skills/worker-management/references/worker-openclaw.json.tmpl b/manager/agent/skills/worker-management/references/worker-openclaw.json.tmpl
index d163b30b..79a31ac1 100644
--- a/manager/agent/skills/worker-management/references/worker-openclaw.json.tmpl
+++ b/manager/agent/skills/worker-management/references/worker-openclaw.json.tmpl
@@ -50,6 +50,7 @@
           { "id": "claude-sonnet-4-6", "name": "claude-sonnet-4-6", "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "claude-haiku-4-5",  "name": "claude-haiku-4-5",  "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "qwen3.5-plus",      "name": "qwen3.5-plus",      "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
+          { "id": "qwen3.6-plus",      "name": "qwen3.6-plus",      "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "deepseek-chat",     "name": "deepseek-chat",     "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
           { "id": "deepseek-reasoner", "name": "deepseek-reasoner", "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
           { "id": "kimi-k2.5",         "name": "kimi-k2.5",         "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text", "image"] },
@@ -78,6 +79,7 @@
         "hiclaw-gateway/claude-sonnet-4-6": { "alias": "claude-sonnet-4-6" },
         "hiclaw-gateway/claude-haiku-4-5":  { "alias": "claude-haiku-4-5" },
         "hiclaw-gateway/qwen3.5-plus":      { "alias": "qwen3.5-plus" },
+        "hiclaw-gateway/qwen3.6-plus":      { "alias": "qwen3.6-plus" },
         "hiclaw-gateway/deepseek-chat":     { "alias": "deepseek-chat" },
         "hiclaw-gateway/deepseek-reasoner": { "alias": "deepseek-reasoner" },
         "hiclaw-gateway/kimi-k2.5":         { "alias": "kimi-k2.5" },
diff --git a/manager/agent/skills/worker-management/scripts/generate-worker-config.sh b/manager/agent/skills/worker-management/scripts/generate-worker-config.sh
index 5f5d15de..6daed1c7 100644
--- a/manager/agent/skills/worker-management/scripts/generate-worker-config.sh
+++ b/manager/agent/skills/worker-management/scripts/generate-worker-config.sh
@@ -16,7 +16,7 @@ source /opt/hiclaw/scripts/lib/hiclaw-env.sh
 WORKER_NAME="$1"
 WORKER_MATRIX_TOKEN="$2"
 WORKER_GATEWAY_KEY="$3"
-MODEL_NAME="${4:-${HICLAW_DEFAULT_MODEL:-qwen3.5-plus}}"
+MODEL_NAME="${4:-${HICLAW_DEFAULT_MODEL:-qwen3.6-plus}}"
 TEAM_LEADER_NAME="${5:-}"
 # Strip provider prefix if caller passed "hiclaw-gateway/<model>" by mistake
 MODEL_NAME="${MODEL_NAME#hiclaw-gateway/}"
@@ -46,6 +46,8 @@ case "${MODEL_NAME}" in
         CTX=200000; MAX=64000 ;;
     qwen3.5-plus)
         CTX=200000; MAX=64000 ;;
+    qwen3.6-plus)
+        CTX=1000000; MAX=64000 ;;
     deepseek-chat|deepseek-reasoner|kimi-k2.5)
         CTX=256000; MAX=128000 ;;
     glm-5|MiniMax-M2.7|MiniMax-M2.7-highspeed|MiniMax-M2.5)
@@ -60,7 +62,7 @@ esac
 
 # Resolve input modalities: only vision-capable models get "image"
 case "${MODEL_NAME}" in
-    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|kimi-k2.5)
+    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|qwen3.6-plus|kimi-k2.5)
         INPUT='["text", "image"]' ;;
     *)
         INPUT='["text"]' ;;
diff --git a/manager/agent/skills/worker-model-switch/SKILL.md b/manager/agent/skills/worker-model-switch/SKILL.md
index 9c5b4d16..68fc5f13 100644
--- a/manager/agent/skills/worker-model-switch/SKILL.md
+++ b/manager/agent/skills/worker-model-switch/SKILL.md
@@ -91,5 +91,6 @@ When the human admin requests switching a Worker to a model you don't recognize,
 | claude-sonnet-4-6 | 1,000,000 | 64,000 |
 | claude-haiku-4-5 | 200,000 | 64,000 |
 | qwen3.5-plus | 200,000 | 64,000 |
+| qwen3.6-plus | 1,000,000 | 64,000 |
 | deepseek-chat / deepseek-reasoner / kimi-k2.5 | 256,000 | 128,000 |
 | glm-5 / MiniMax-M2.7 / MiniMax-M2.7-highspeed / MiniMax-M2.5 | 200,000 | 128,000 |
diff --git a/manager/agent/skills/worker-model-switch/scripts/update-worker-model.sh b/manager/agent/skills/worker-model-switch/scripts/update-worker-model.sh
index 628bfc25..83104050 100644
--- a/manager/agent/skills/worker-model-switch/scripts/update-worker-model.sh
+++ b/manager/agent/skills/worker-model-switch/scripts/update-worker-model.sh
@@ -41,6 +41,8 @@ _resolve_model_params() {
             CTX=200000; MAX=64000 ;;
         qwen3.5-plus)
             CTX=200000; MAX=64000 ;;
+        qwen3.6-plus)
+            CTX=1000000; MAX=64000 ;;
         deepseek-chat|deepseek-reasoner|kimi-k2.5)
             CTX=256000; MAX=128000 ;;
         glm-5|MiniMax-M2.7|MiniMax-M2.7-highspeed|MiniMax-M2.5)
@@ -54,7 +56,7 @@ _resolve_model_params() {
     fi
     # Resolve input modalities: only vision-capable models get "image"
     case "${model}" in
-        gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|kimi-k2.5)
+        gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|qwen3.6-plus|kimi-k2.5)
             INPUT='["text", "image"]' ;;
         *)
             INPUT='["text"]' ;;
diff --git a/manager/configs/known-models.json b/manager/configs/known-models.json
index fac462a3..742c5370 100644
--- a/manager/configs/known-models.json
+++ b/manager/configs/known-models.json
@@ -7,6 +7,7 @@
   { "id": "claude-sonnet-4-6", "name": "claude-sonnet-4-6", "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
   { "id": "claude-haiku-4-5",  "name": "claude-haiku-4-5",  "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
   { "id": "qwen3.5-plus",      "name": "qwen3.5-plus",      "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
+  { "id": "qwen3.6-plus",      "name": "qwen3.6-plus",      "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
   { "id": "deepseek-chat",     "name": "deepseek-chat",     "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
   { "id": "deepseek-reasoner", "name": "deepseek-reasoner", "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
   { "id": "kimi-k2.5",         "name": "kimi-k2.5",         "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text", "image"] },
diff --git a/manager/configs/manager-openclaw.json.tmpl b/manager/configs/manager-openclaw.json.tmpl
index b3f86697..52194e5a 100644
--- a/manager/configs/manager-openclaw.json.tmpl
+++ b/manager/configs/manager-openclaw.json.tmpl
@@ -51,6 +51,7 @@
           { "id": "claude-sonnet-4-6", "name": "claude-sonnet-4-6", "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "claude-haiku-4-5",  "name": "claude-haiku-4-5",  "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "qwen3.5-plus",      "name": "qwen3.5-plus",      "reasoning": true,  "contextWindow": 200000,  "maxTokens": 64000,  "input": ["text", "image"] },
+          { "id": "qwen3.6-plus",      "name": "qwen3.6-plus",      "reasoning": true,  "contextWindow": 1000000, "maxTokens": 64000,  "input": ["text", "image"] },
           { "id": "deepseek-chat",     "name": "deepseek-chat",     "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
           { "id": "deepseek-reasoner", "name": "deepseek-reasoner", "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text"] },
           { "id": "kimi-k2.5",         "name": "kimi-k2.5",         "reasoning": true,  "contextWindow": 256000,  "maxTokens": 128000, "input": ["text", "image"] },
@@ -79,6 +80,7 @@
         "hiclaw-gateway/claude-sonnet-4-6": { "alias": "claude-sonnet-4-6" },
         "hiclaw-gateway/claude-haiku-4-5":  { "alias": "claude-haiku-4-5" },
         "hiclaw-gateway/qwen3.5-plus":      { "alias": "qwen3.5-plus" },
+        "hiclaw-gateway/qwen3.6-plus":      { "alias": "qwen3.6-plus" },
         "hiclaw-gateway/deepseek-chat":     { "alias": "deepseek-chat" },
         "hiclaw-gateway/deepseek-reasoner": { "alias": "deepseek-reasoner" },
         "hiclaw-gateway/kimi-k2.5":         { "alias": "kimi-k2.5" },
diff --git a/manager/scripts/init/start-manager-agent.sh b/manager/scripts/init/start-manager-agent.sh
index 4ade2694..84c6f65c 100755
--- a/manager/scripts/init/start-manager-agent.sh
+++ b/manager/scripts/init/start-manager-agent.sh
@@ -454,7 +454,7 @@ log "Generating Manager openclaw.json..."
 export MANAGER_MATRIX_TOKEN="${MANAGER_TOKEN}"
 export MANAGER_GATEWAY_KEY="${HICLAW_MANAGER_GATEWAY_KEY}"
 # Resolve model parameters based on model name
-MODEL_NAME="${HICLAW_DEFAULT_MODEL:-qwen3.5-plus}"
+MODEL_NAME="${HICLAW_DEFAULT_MODEL:-qwen3.6-plus}"
 case "${MODEL_NAME}" in
     gpt-5.3-codex|gpt-5-mini|gpt-5-nano)
         export MODEL_CONTEXT_WINDOW=400000 MODEL_MAX_TOKENS=128000 ;;
@@ -466,6 +466,8 @@ case "${MODEL_NAME}" in
         export MODEL_CONTEXT_WINDOW=200000 MODEL_MAX_TOKENS=64000 ;;
     qwen3.5-plus)
         export MODEL_CONTEXT_WINDOW=200000 MODEL_MAX_TOKENS=64000 ;;
+    qwen3.6-plus)
+        export MODEL_CONTEXT_WINDOW=1000000 MODEL_MAX_TOKENS=64000 ;;
     deepseek-chat|deepseek-reasoner|kimi-k2.5)
         export MODEL_CONTEXT_WINDOW=256000 MODEL_MAX_TOKENS=128000 ;;
     glm-5|MiniMax-M2.7|MiniMax-M2.7-highspeed|MiniMax-M2.5)
@@ -490,7 +492,7 @@ log "Matrix E2EE: ${MATRIX_E2EE_ENABLED}"
 
 # Resolve input modalities: only vision-capable models get "image"
 case "${MODEL_NAME}" in
-    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|kimi-k2.5)
+    gpt-5.4|gpt-5.3-codex|gpt-5-mini|gpt-5-nano|claude-opus-4-6|claude-sonnet-4-6|claude-haiku-4-5|qwen3.5-plus|qwen3.6-plus|kimi-k2.5)
         export MODEL_INPUT='["text", "image"]' ;;
     *)
         export MODEL_INPUT='["text"]' ;;
diff --git a/tests/test-15-import-worker-zip.sh b/tests/test-15-import-worker-zip.sh
index 6916a04f..44f9b0b9 100644
--- a/tests/test-15-import-worker-zip.sh
+++ b/tests/test-15-import-worker-zip.sh
@@ -82,7 +82,7 @@ exec_in_manager bash -c "
   \"version\": 1,
   \"worker\": {
     \"suggested_name\": \"${TEST_WORKER}\",
-    \"model\": \"qwen3.5-plus\"
+    \"model\": \"qwen3.6-plus\"
   },
   \"source\": {
     \"hostname\": \"integration-test\"
diff --git a/tests/test-17-worker-config-verify.sh b/tests/test-17-worker-config-verify.sh
index 8cfc1df1..ea6d5e51 100644
--- a/tests/test-17-worker-config-verify.sh
+++ b/tests/test-17-worker-config-verify.sh
@@ -46,7 +46,7 @@ exec_in_manager bash -c "
   \"version\": 1,
   \"worker\": {
     \"suggested_name\": \"${TEST_WORKER}\",
-    \"model\": \"qwen3.5-plus\"
+    \"model\": \"qwen3.6-plus\"
   }
 }
 MANIFEST
diff --git a/tests/test-20-inline-worker-config.sh b/tests/test-20-inline-worker-config.sh
index fea7800f..d7fc4b38 100755
--- a/tests/test-20-inline-worker-config.sh
+++ b/tests/test-20-inline-worker-config.sh
@@ -83,7 +83,7 @@ kind: Worker
 metadata:
   name: ${TEST_WORKER}
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
   soul: |
 $(echo "${SOUL_CONTENT}" | sed 's/^/    /')
   agents: |
@@ -259,7 +259,7 @@ exec_in_manager bash -c "
   \"version\": 1,
   \"worker\": {
     \"suggested_name\": \"${TEST_WORKER_OVERRIDE}\",
-    \"model\": \"qwen3.5-plus\"
+    \"model\": \"qwen3.6-plus\"
   },
   \"source\": {
     \"hostname\": \"integration-test\"
@@ -312,7 +312,7 @@ kind: Worker
 metadata:
   name: ${TEST_WORKER_OVERRIDE}
 spec:
-  model: qwen3.5-plus
+  model: qwen3.6-plus
   package: ${PKG_URI}
   soul: |
 $(echo "${OVERRIDE_SOUL}" | sed 's/^/    /')