feat: add support for custom reasoning models

[mrT23]

User description

#1540

---

PR Type

Enhancement, Documentation

---

Description

• Added support for custom reasoning models in AI handler.

• Updated logic to handle models without temperature or chat-style inputs.

• Documented configuration for custom reasoning models in usage guide.

• Introduced a new configuration option custom_reasoning_model.

---

Changes walkthrough 📝

	Relevant files
Enhancement	litellm_ai_handler.py Enhance AI handler to support custom reasoning models        pr_agent/algo/ai_handlers/litellm_ai_handler.py Modified logic to combine system and user prompts for custom reasoning models. Updated temperature handling to accommodate custom reasoning models. +2/-2
Documentation	changing_a_model.md Update usage guide for custom reasoning models                      docs/docs/usage-guide/changing_a_model.md Added documentation for custom reasoning models. Explained limitations of reasoning models regarding temperature and chat-style inputs. +3/-0
Configuration changes	configuration.toml Add configuration for custom reasoning models                        pr_agent/settings/configuration.toml Introduced custom_reasoning_model configuration option. Documented its purpose and default value. +1/-0

---

Need help?

Type /help how to ... in the comments thread for any questions about Qodo Merge usage.

Check out the documentation for more information.

[qodo-merge-pro-for-open-source]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

🎫 Ticket compliance analysis ✅ 1540 - Fully compliant Compliant requirements: Handle Azure OpenAI Service models that don't support temperature parameter Fix error when using o3-mini model with temperature parameter

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Logic Validation The condition for skipping temperature parameter should be verified to ensure it correctly identifies models that don't support temperature if model not in self.no_support_temperature_models or get_settings().config.custom_reasoning_model: kwargs["temperature"] = temperature

[qodo-merge-pro-for-open-source]

PR Code Suggestions ✨

Latest suggestions up to 35059ca

Category	Suggestion	Impact
Possible issue	Add null check for config Add a check to verify that get_settings().config exists before accessing custom_reasoning_model to prevent potential NullPointerException pr_agent/algo/ai_handlers/litellm_ai_handler.py [208] -if model in self.user_message_only_models or get_settings().config.custom_reasoning_model: +if model in self.user_message_only_models or (hasattr(get_settings(), 'config') and get_settings().config.custom_reasoning_model): Apply this suggestion Suggestion importance[1-10]: 7 __ Why: The suggestion adds important defensive programming by checking for the existence of the config attribute before accessing it, which could prevent runtime errors if the settings object is not properly initialized.	Medium
	Add consistent null check Apply the same null check pattern for the second occurrence of config access to maintain consistency and prevent potential crashes pr_agent/algo/ai_handlers/litellm_ai_handler.py [230] -if model not in self.no_support_temperature_models and not get_settings().config.custom_reasoning_model: +if model not in self.no_support_temperature_models and not (hasattr(get_settings(), 'config') and get_settings().config.custom_reasoning_model): Apply this suggestion Suggestion importance[1-10]: 7 __ Why: This suggestion maintains consistency with the first fix and adds the same defensive programming pattern to prevent potential null pointer exceptions in another similar code location.	Medium
More

---

Previous suggestions

✅ Suggestions up to commit 0317951

Category	Suggestion	Impact
Possible issue	✅ Fix incorrect temperature handling logic The condition for handling temperature settings should be inverted - if a model doesn't support temperature or is a custom reasoning model, temperature should be excluded, not included. pr_agent/algo/ai_handlers/litellm_ai_handler.py [230-231] -if model not in self.no_support_temperature_models or get_settings().config.custom_reasoning_model: +if model not in self.no_support_temperature_models and not get_settings().config.custom_reasoning_model: kwargs["temperature"] = temperature [Suggestion has been applied] Suggestion importance[1-10]: 10 __ Why: The current logic is incorrect and would add temperature to models that don't support it. The suggested fix correctly excludes temperature for both unsupported models and custom reasoning models, preventing potential API errors.	High

✅ Suggestions up to commit 4edb8b8

Category	Suggestion	Impact
General	Add warning for custom model settings Add a warning log when custom_reasoning_model is enabled to track when system messages and temperature settings are being bypassed. pr_agent/algo/ai_handlers/litellm_ai_handler.py [208-212] if model in self.user_message_only_models or get_settings().config.custom_reasoning_model: + if get_settings().config.custom_reasoning_model: + get_logger().warning("Custom reasoning model enabled: bypassing system message and temperature settings") user = f"{system}\n\n\n{user}" system = "" get_logger().info(f"Using model {model}, combining system and user prompts") messages = [{"role": "user", "content": user}] Suggestion importance[1-10]: 7 __ Why: Adding a warning log when custom_reasoning_model is enabled provides better observability and helps track when important model settings are being bypassed, making it easier to debug and monitor system behavior.	Medium
	✅ Clarify configuration comment documentation Improve the configuration comment to clearly explain the implications of enabling custom_reasoning_model. pr_agent/settings/configuration.toml [20] -custom_reasoning_model = false # when true, no system message and not temperature will be used +custom_reasoning_model = false # when true, disables system messages and temperature controls for models that don't support chat-style inputs [Suggestion has been applied] Suggestion importance[1-10]: 6 __ Why: The improved comment provides clearer and more precise documentation about the implications of enabling custom_reasoning_model, helping users better understand its impact on chat-style inputs.	Low