RooCodeInc · hannesrudolph · Dec 19, 2025 · Dec 19, 2025
@@ -24,6 +24,7 @@ If you want to live on the edge and try things out before it's released, we have
 
 ### Version 3.36
 
+*   [3.36.15](/update-notes/v3.36.15) (2025-12-19)
 *   [3.36.14](/update-notes/v3.36.14) (2025-12-18)
 *   [3.36.13](/update-notes/v3.36.13) (2025-12-18)
 *   [3.36.12](/update-notes/v3.36.12) (2025-12-18)

@@ -0,0 +1,36 @@
+---
+description: Roo Code 3.36.15 adds a 1M context window option for Claude Sonnet 4.5 on Vertex AI, improves chat error troubleshooting, and expands native tool calling support.
+keywords:
+  - roo code 3.36.15
+  - vertex ai
+  - 1m context
+  - native tool calling
+  - error diagnostics
+  - bug fixes
+image: /img/v3.36.15/v3.36.15.png
+---
+
+# Roo Code 3.36.15 Release Notes (2025-12-19)
+
+This release adds a 1M context window option for Claude Sonnet 4.5 on Vertex AI, improves chat error troubleshooting, and expands native tool calling support.
+
+<img src="/img/v3.36.15/v3.36.15.png" alt="Roo Code v3.36.15 Release" width="600" />
+
+## Vertex AI: 1M context window for Claude Sonnet 4.5
+
+When you use **Claude Sonnet 4.5** on **Vertex AI**, you can now enable a **1M context window** option for supported models ([#10209](https://github.com/RooCodeInc/Roo-Code/pull/10209)).
+
+## Chat error troubleshooting improvements
+
+Chat error states now make it easier to understand what went wrong and to share the right details when filing a bug report:
+
+* **Clearer error visibility**: Error rows more consistently surface full error details (including status codes) via a more obvious **View details** affordance ([#10204](https://github.com/RooCodeInc/Roo-Code/pull/10204))
+* **Downloadable diagnostics**: You can generate a local diagnostics file from a chat error (including error metadata and the API conversation history) so you can review/redact and share it with an issue report ([#10188](https://github.com/RooCodeInc/Roo-Code/pull/10188))
+
+## Bug Fixes
+
+* **Native tool calling support for LM Studio and Qwen-Code**: Fixes an issue where these providers were missing OpenAI-style native tool call support, which could make tool use unreliable compared to other providers ([#10208](https://github.com/RooCodeInc/Roo-Code/pull/10208))
+* **More reliable tool defaults for OpenAI Compatible providers**: Fixes cases where tool calling could be inconsistent unless you manually adjusted custom model info, by applying native tool defaults unless you’ve explicitly overridden them ([#10213](https://github.com/RooCodeInc/Roo-Code/pull/10213))
+* **Requesty native tool calls enabled**: Fixes native tool calling defaults for the Requesty provider (and aligns behavior for Unbound) so tool use is more consistent, especially when model metadata is cached ([#10211](https://github.com/RooCodeInc/Roo-Code/pull/10211))
+* **Strict JSON Schema compatibility**: Fixes an issue where some MCP tool schemas could fail strict validation due to missing `additionalProperties: false` on object schemas ([#10210](https://github.com/RooCodeInc/Roo-Code/pull/10210))
+* **Refresh models cache reliability**: Fixes an issue where **Refresh models** could fail to fully flush/refresh cached model lists for some providers, and improves correctness of initial model selection when starting a new task ([#9870](https://github.com/RooCodeInc/Roo-Code/pull/9870))
@@ -59,6 +59,17 @@ The DeepSeek provider's `deepseek-reasoner` model now supports "interleaved thin
 
 Models that support native tool calling now default to using native protocol instead of XML. The XML protocol is still available in provider settings ([#10186](https://github.com/RooCodeInc/Roo-Code/pull/10186)).
 
+## Vertex AI: 1M context window for Claude Sonnet 4.5
+
+When you use **Claude Sonnet 4.5** on **Vertex AI**, you can enable a **1M context window** option for supported models ([#10209](https://github.com/RooCodeInc/Roo-Code/pull/10209)).
+
+## Chat error troubleshooting improvements
+
+Chat error states now make it easier to understand what went wrong and to share the right details when filing a bug report:
+
+* **Clearer error visibility**: Error rows more consistently surface full error details (including status codes) via a more obvious **View details** affordance ([#10204](https://github.com/RooCodeInc/Roo-Code/pull/10204))
+* **Downloadable diagnostics**: You can generate a local diagnostics file from a chat error (including error metadata and the API conversation history) so you can review/redact and share it with an issue report ([#10188](https://github.com/RooCodeInc/Roo-Code/pull/10188))
+
 ## QOL Improvements
 
 * **Symlink support for slash commands**: Share and organize commands across projects using symlinks for individual files or directories, with command names derived from symlink names ([#9838](https://github.com/RooCodeInc/Roo-Code/pull/9838))
@@ -132,6 +143,11 @@ Models that support native tool calling now default to using native protocol ins
 * **Task Resumption**: Tasks no longer break when resuming after changing the Native Tool Calling setting ([#10192](https://github.com/RooCodeInc/Roo-Code/pull/10192))
 * **Bedrock Embedder CloudTrail Fix**: AWS Bedrock users now see Roo Code identified in CloudTrail logs when using Codebase Indexing. (thanks jackrein!) ([#10166](https://github.com/RooCodeInc/Roo-Code/pull/10166))
 * **MCP Compatibility with OpenAI Providers**: Fixes an issue where MCP servers using `format: "uri"` in their tool schemas would fail with OpenAI providers ([#10198](https://github.com/RooCodeInc/Roo-Code/pull/10198))
+* **Native tool calling support for LM Studio and Qwen-Code**: Fixes an issue where these providers were missing OpenAI-style native tool call support, which could make tool use unreliable compared to other providers ([#10208](https://github.com/RooCodeInc/Roo-Code/pull/10208))
+* **More reliable tool defaults for OpenAI Compatible providers**: Fixes cases where tool calling could be inconsistent unless you manually adjusted custom model info, by applying native tool defaults unless you've explicitly overridden them ([#10213](https://github.com/RooCodeInc/Roo-Code/pull/10213))
+* **Requesty native tool calls enabled**: Fixes native tool calling defaults for the Requesty provider (and aligns behavior for Unbound) so tool use is more consistent, especially when model metadata is cached ([#10211](https://github.com/RooCodeInc/Roo-Code/pull/10211))
+* **MCP tool schemas work with stricter validation**: Fixes an issue where some MCP tool schemas could fail strict validation due to missing `additionalProperties: false` on object schemas ([#10210](https://github.com/RooCodeInc/Roo-Code/pull/10210))
+* **Refresh models cache reliability**: Fixes an issue where **Refresh models** could fail to fully flush/refresh cached model lists for some providers, and improves correctness of initial model selection when starting a new task ([#9870](https://github.com/RooCodeInc/Roo-Code/pull/9870))
 
 ## Misc Improvements
 

@@ -173,6 +173,7 @@ const sidebars: SidebarsConfig = {
           label: '3.36',
           items: [
             { type: 'doc', id: 'update-notes/v3.36', label: '3.36 Combined' },
+            { type: 'doc', id: 'update-notes/v3.36.15', label: '3.36.15' },
             { type: 'doc', id: 'update-notes/v3.36.14', label: '3.36.14' },
             { type: 'doc', id: 'update-notes/v3.36.13', label: '3.36.13' },
             { type: 'doc', id: 'update-notes/v3.36.12', label: '3.36.12' },