MINA LABS
Start creating free
Ollama has released version 0.34.0, introducing direct support for using Ollama models within ChatGPT Desktop. This allows users to leverage open-source models through an existing workflow without switching between applications. The feature is available via setup options in the Ollama app on macOS, as detailed in the release notes. Alongside this integration, the update enhances structured output performance on Apple Silicon chips and introduces compatibility with OpenAI-style client tools for search and response compaction.
For developers and creators who rely on local or self-hosted models, this update simplifies workflows by bridging Ollama with ChatGPT Desktop’s interface. Previously, users might have needed to manually route requests or use separate tools to access Ollama models within ChatGPT’s ecosystem. The integration reduces friction for those who prefer ChatGPT’s UI but want the flexibility of open models like Llama or Mistral. This could streamline tasks such as code generation, content drafting, or data analysis without abandoning familiar tools.
The release also addresses performance bottlenecks on Apple Silicon devices, which are increasingly central to creative and development workflows. Structured output—critical for tasks like JSON parsing or code generation—now executes faster on M1/M2/M3 chips. Additionally, the OpenAI-compatible client tools enable developers to implement features like dynamic search or response summarization using Ollama’s backend. This aligns with growing demand for hybrid setups where open models coexist with proprietary APIs.
Ollama v0.34.0 is available for download on macOS, Windows, and Linux through the official website. For those using Mina Labs’ infrastructure, the model can be accessed at a cost of 8 per image generation. Edits or modifications to generated content will cost 8 per image. These pricing tokens reflect the current rate structure for leveraging Ollama models through Mina Labs’ platform, which prioritizes open-source compatibility and decentralized execution.
Use cases for this update span technical and creative domains. Developers can now test prompts in ChatGPT Desktop while routing computations to local Ollama models, reducing reliance on cloud APIs. Designers and writers might benefit from the improved Apple Silicon performance when generating or refining visual or textual content. The OpenAI tool compatibility also opens doors for integrating Ollama models into existing applications that rely on OpenAI’s SDK, expanding their flexibility without rewriting code.
Beyond the headline features, the update fixes edge cases where images failed to render correctly in compacted responses—a common issue when optimizing for shorter, more efficient outputs. This ensures that visual elements remain intact even when responses are condensed for clarity or brevity. The full changelog, available here, details other minor improvements and bug fixes.
For teams managing local AI deployments, this release simplifies coordination between different tools. A developer working on a macOS workstation can now prototype ideas in ChatGPT Desktop while ensuring all model computations stay on their machine, avoiding data exposure risks. Similarly, educators or researchers can demonstrate open models using ChatGPT’s polished interface while maintaining control over their infrastructure.
While the update focuses on integration and performance, it also signals Ollama’s broader strategy of making open models more accessible without sacrificing control. By aligning with popular interfaces like ChatGPT Desktop, the project lowers the barrier for users hesitant to adopt open-source tools. This could accelerate experimentation with self-hosted models in environments where privacy, cost, or customization matter more than proprietary alternatives.
In short, Ollama v0.34.0 is a pragmatic update that bridges gaps between user-friendly interfaces and open-source flexibility. Whether you’re optimizing workflows, improving performance, or exploring hybrid setups, the release provides tangible benefits without requiring a shift in existing practices.