Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions docs/update-notes/index.md
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,8 @@ If you want to live on the edge and try things out before it's released, we have
### Version 3.37

* [3.37](/update-notes/v3.37) (2025-12-22)
* [3.37.0](/update-notes/v3.37.0) (2025-12-22)
* [3.37.1](/update-notes/v3.37.1) (2025-12-23)

---

Expand Down
58 changes: 58 additions & 0 deletions docs/update-notes/v3.37.0.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
---
description: Roo Code 3.37 introduces custom tool calling, expands provider support, and improves tool-call reliability.
keywords:
- roo code 3.37.0
- new features
- bug fixes
image: /img/v3.37.0/v3.37.0.png
---

# Roo Code 3.37.0 Release Notes (2025-12-22)

Roo Code 3.37 introduces experimental custom tools, adds new provider capabilities, and improves tool-call reliability.

<img src="/img/v3.37.0/v3.37.0.png" alt="Roo Code v3.37.0 Release" width="600" />

## New models
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In docs/update-notes/v3.37.0.mdx the ## New models section is immediately followed by another ## ... heading, which makes the “New models” section empty and flattens the hierarchy. Consider making the Z.ai and MiniMax headings ### under “New models” to match the intended structure.

Fix it with Roo Code or mention @roomote and request a fix.


## Z.ai GLM-4.7 (thinking mode)

GLM-4.7 is now available directly through the Z.ai provider in Roo Code, as well as via the Roo Code Cloud provider (and other provider routes that surface Z.ai). It’s a strong coding model for agentic workflows, with improved multilingual coding, terminal tasks, tool use, and complex reasoning compared to GLM-4.6 ([#10282](https://github.com/RooCodeInc/Roo-Code/pull/10282)).

<img src="/img/v3.37/v3.37-2.png" alt="GLM-4.7 benchmark performance" width="600" />

## MiniMax M2.1 improvements

MiniMax M2.1 is now available directly through the MiniMax provider in Roo Code, as well as via the Roo Code Cloud provider (and other provider routes that surface MiniMax). It’s a strong pick for agentic coding workflows, with better tool use, instruction following, and long-horizon planning for multi-step tasks—and it’s fast ([#10284](https://github.com/RooCodeInc/Roo-Code/pull/10284)).

<img src="/img/v3.37/v3.37-1.png" alt="MiniMax M2.1 Benchmark Performance" width="600" />

## Experimental custom tools

You can now define and use **custom tools** so Roo can call your project- or team-specific actions like built-in tools. This makes it easier to standardize workflows across a team by shipping tool schemas alongside your project, instead of repeatedly re-prompting the same steps ([#10083](https://github.com/RooCodeInc/Roo-Code/pull/10083)).

## Bug Fixes

* Fixes an issue where Roo could appear stuck after a tool call with some OpenAI-compatible providers when streaming ended at the tool-calls boundary (thanks torxeon!) ([#10280](https://github.com/RooCodeInc/Roo-Code/pull/10280))
* Fixes an issue where Roo could appear stuck after a tool call with some OpenAI-compatible providers by ensuring final tool-call completion events are emitted ([#10293](https://github.com/RooCodeInc/Roo-Code/pull/10293))
* Fixes an issue where MCP tools could break under strict schema mode when optional parameters were treated as required ([#10220](https://github.com/RooCodeInc/Roo-Code/pull/10220))
* Fixes an issue where the built-in `read_file` tool could fail on some models due to invalid schema normalization for optional array parameters ([#10276](https://github.com/RooCodeInc/Roo-Code/pull/10276))
* Fixes an issue where `search_replace` / `search_and_replace` could miss matches on CRLF files, improving cross-platform search-and-replace reliability ([#10288](https://github.com/RooCodeInc/Roo-Code/pull/10288))
* Fixes an issue where Requesty’s **Refresh Models** could leave the model list stale by not including credentials in the refresh flow (thanks requesty-JohnCosta27!) ([#10273](https://github.com/RooCodeInc/Roo-Code/pull/10273))
* Fixes an issue where Chutes model loading could fail if the provider returned malformed model entries ([#10279](https://github.com/RooCodeInc/Roo-Code/pull/10279))
* Fixes an issue where `reasoning_details` could be merged/ordered incorrectly during streaming, improving reliability for providers that depend on strict reasoning serialization ([#10285](https://github.com/RooCodeInc/Roo-Code/pull/10285))
* Fixes an issue where DeepSeek-reasoner could error after condensation if the condensed summary lacked required reasoning fields ([#10292](https://github.com/RooCodeInc/Roo-Code/pull/10292))

## Misc Improvements

* **Cleaner eval logs**: Deduplicates repetitive message log entries so eval traces are easier to read ([#10286](https://github.com/RooCodeInc/Roo-Code/pull/10286))

## QOL Improvements

* New tasks now default to native tool calling on models that support it, reducing the need for manual tool protocol selection ([#10281](https://github.com/RooCodeInc/Roo-Code/pull/10281))

## Provider Updates

* Improves Z.ai thinking model message formatting by attaching `environment_details` to tool results instead of emitting separate system messages ([#10289](https://github.com/RooCodeInc/Roo-Code/pull/10289))
* LiteLLM no longer sends `parallel_tool_calls`, improving tool-call compatibility (thanks farazoman!) ([#10274](https://github.com/RooCodeInc/Roo-Code/pull/10274))

31 changes: 31 additions & 0 deletions docs/update-notes/v3.37.1.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
---
description: Roo Code 3.37.1 refreshes the signup flow, improves OpenAI-compatible tool calling, and fixes several reliability issues.
keywords:
- roo code 3.37.1
- signup
- openai compatible
- bug fixes
image: /img/v3.37.1/v3.37.1.png
---

# Roo Code 3.37.1 Release Notes (2025-12-23)

This release refreshes the signup flow, improves tool-calling reliability for OpenAI-compatible providers, and fixes several chat reliability issues.

<img src="/img/v3.37.1/v3.37.1.png" alt="Roo Code v3.37.1 Release" width="600" />

## QOL Improvements

* Improves tool-calling reliability for Roo Code Cloud by preventing tool-result metadata (like `environment_details`) from interrupting tool call sequences ([#10301](https://github.com/RooCodeInc/Roo-Code/pull/10301))
* Improves tool-calling reliability across OpenAI-compatible providers by merging trailing tool-result text into the last tool message, reducing cases where tool call sequences get interrupted ([#10299](https://github.com/RooCodeInc/Roo-Code/pull/10299))

## Bug Fixes

* Fixes an issue where Roo could show errors when a provider returned an empty assistant message by retrying once and only showing an error if the problem repeats ([#10297](https://github.com/RooCodeInc/Roo-Code/pull/10297))
* Fixes an issue where OpenAI/OpenAI-compatible chats could fail to use native tools when custom model info didn’t explicitly set tool support, by sending native tool definitions by default ([#10314](https://github.com/RooCodeInc/Roo-Code/pull/10314))
* Fixes an issue where Roo could send malformed `reasoning_details` data after transforming conversation history, preventing provider-side errors and improving compatibility with OpenAI Responses-style reasoning blocks ([#10313](https://github.com/RooCodeInc/Roo-Code/pull/10313))
* Fixes an issue where “ask” flows could hang if your reply was queued instead of being delivered as an ask response, so conversations continue reliably ([#10315](https://github.com/RooCodeInc/Roo-Code/pull/10315))

## Misc Improvements

* **Provider-centric signup**: Makes Roo easier to get started with by defaulting the welcome/sign-up flow to the Roo provider (while keeping other providers available), so you can reach a working setup faster with fewer choices up front ([#10306](https://github.com/RooCodeInc/Roo-Code/pull/10306))
7 changes: 7 additions & 0 deletions docs/update-notes/v3.37.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -42,14 +42,21 @@ You can now define and use **custom tools** so Roo can call your project- or tea
* Fixes an issue where Chutes model loading could fail if the provider returned malformed model entries ([#10279](https://github.com/RooCodeInc/Roo-Code/pull/10279))
* Fixes an issue where `reasoning_details` could be merged/ordered incorrectly during streaming, improving reliability for providers that depend on strict reasoning serialization ([#10285](https://github.com/RooCodeInc/Roo-Code/pull/10285))
* Fixes an issue where DeepSeek-reasoner could error after condensation if the condensed summary lacked required reasoning fields ([#10292](https://github.com/RooCodeInc/Roo-Code/pull/10292))
* Fixes an issue where Roo could show errors when a provider returned an empty assistant message by retrying once and only showing an error if the problem repeats ([#10297](https://github.com/RooCodeInc/Roo-Code/pull/10297))
* Fixes an issue where OpenAI/OpenAI-compatible chats could fail to use native tools when custom model info didn’t explicitly set tool support, by sending native tool definitions by default ([#10314](https://github.com/RooCodeInc/Roo-Code/pull/10314))
* Fixes an issue where Roo could send malformed `reasoning_details` data after transforming conversation history, preventing provider-side errors and improving compatibility with OpenAI Responses-style reasoning blocks ([#10313](https://github.com/RooCodeInc/Roo-Code/pull/10313))
* Fixes an issue where “ask” flows could hang if your reply was queued instead of being delivered as an ask response, so conversations continue reliably ([#10315](https://github.com/RooCodeInc/Roo-Code/pull/10315))

## Misc Improvements

* **Cleaner eval logs**: Deduplicates repetitive message log entries so eval traces are easier to read ([#10286](https://github.com/RooCodeInc/Roo-Code/pull/10286))
* **Provider-centric signup**: Makes Roo easier to get started with by defaulting the welcome/sign-up flow to the Roo provider (while keeping other providers available), so you can reach a working setup faster with fewer choices up front ([#10306](https://github.com/RooCodeInc/Roo-Code/pull/10306))

## QOL Improvements

* New tasks now default to native tool calling on models that support it, reducing the need for manual tool protocol selection ([#10281](https://github.com/RooCodeInc/Roo-Code/pull/10281))
* Improves tool-calling reliability for Roo Code Cloud by preventing tool-result metadata (like `environment_details`) from interrupting tool call sequences ([#10301](https://github.com/RooCodeInc/Roo-Code/pull/10301))
* Improves tool-calling reliability across OpenAI-compatible providers by merging trailing tool-result text into the last tool message, reducing cases where tool call sequences get interrupted ([#10299](https://github.com/RooCodeInc/Roo-Code/pull/10299))

## Provider Updates

Expand Down
2 changes: 2 additions & 0 deletions sidebars.ts
Original file line number Diff line number Diff line change
Expand Up @@ -176,6 +176,8 @@ const sidebars: SidebarsConfig = {
label: '3.37',
items: [
{ type: 'doc', id: 'update-notes/v3.37', label: '3.37 Combined' },
{ type: 'doc', id: 'update-notes/v3.37.0', label: '3.37.0' },
{ type: 'doc', id: 'update-notes/v3.37.1', label: '3.37.1' },
],
},
{
Expand Down
Binary file added static/img/v3.37.1/v3.37.1.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.