fix(ollama): fixed messages array for ollama, added gpt-5.2

[waleedlatif1]

Summary

• fixed messages array for ollama, added gpt-5.2

Type of Change

• Bug fix
• New feature

Testing

Tested manually

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
docs	Skipped			Dec 11, 2025 9:42pm

[greptile-apps]

Greptile Overview

Greptile Summary

Fixed a bug in Ollama provider integration where messages were being duplicated, and added support for the new gpt-5.2 model.

Key Changes:

• Fixed message duplication in Ollama by conditionally setting the context field to undefined when valid messages exist (previously context was always populated, causing Ollama to add it as an additional user message even when messages were already provided)
• Added gpt-5.2 model to both OpenAI and Azure OpenAI providers with pricing ($1.75 input, $14.0 output), 400K context window, and support for reasoningEffort and verbosity capabilities

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• Both changes are straightforward and correct. The bug fix prevents message duplication by properly handling the context field based on message validity, and the model addition follows the established pattern with consistent configuration across OpenAI and Azure providers. No logical errors or security issues found.
• No files require special attention

Important Files Changed

File Analysis

Filename	Score	Overview
apps/sim/executor/handlers/agent/agent-handler.ts	5/5	Fixed Ollama duplicate message bug by conditionally setting context field to undefined when valid messages exist
apps/sim/providers/models.ts	5/5	Added gpt-5.2 model definition to OpenAI and Azure OpenAI providers with pricing and capabilities

Sequence Diagram

sequenceDiagram
    participant Agent as AgentBlockHandler
    participant Builder as buildProviderRequest
    participant Provider as executeProviderRequest
    participant Ollama as OllamaProvider
    
    Agent->>Builder: buildProviderRequest(config)
    Builder->>Builder: validateMessages(messages)
    
    alt validMessages = true
        Note over Builder: Messages are valid<br/>Set context = undefined<br/>Set systemPrompt = undefined
        Builder->>Provider: { messages, context: undefined }
        Provider->>Ollama: { messages: [...] }
        Ollama->>Ollama: allMessages.push(...messages)
        Note over Ollama: No duplicate messages
    else validMessages = false
        Note over Builder: Messages invalid<br/>Set context = stringifyJSON(messages)<br/>Use systemPrompt
        Builder->>Provider: { context, systemPrompt }
        Provider->>Ollama: { context: "..." }
        Ollama->>Ollama: allMessages.push({ role: 'user', content: context })
        Note over Ollama: Context added as user message
    end
    
    Ollama-->>Provider: Response
    Provider-->>Agent: Result

Loading

[greptile-apps]

Additional Comments (1)

1. apps/sim/providers/ollama/index.ts, line 110-124 (link)

   logic: removed request.context handling that exists in all other providers - this breaks context passing

   All other providers (OpenAI, Anthropic, Google, etc.) handle request.context by inserting it as a user message between the system prompt and the messages array. Removing this from Ollama breaks parity and will cause context to be ignored when using Ollama.

_{2 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[waleedlatif1]

@greptile

[greptile-apps]

_{2 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}