OpenAI cuts GPT 5.6 Sol costs and begins winding down fine-tuning platform
The artificial intelligence firm has lowered pricing for its latest frontier model while restricting new access to its customisation tools, a move that signals a shift in its developer strategy.
OpenAI has announced a price reduction for its GPT 5.6 Sol model, a move that remains in effect until at least 21 November. The adjustment forms part of the company’s Daybreak programme, which releases new frontier models and updates associated aliases to point to the latest available versions.
Under this framework, the alias daybreak-blue-latest currently points to gpt-5.6-sol, while daybreak-red-latest points to gpt-5.6-cyber. As new models are released through the Daybreak programme, these aliases will be updated to reflect the most recent releases, with pricing adjusted to match the underlying model.
Simultaneously, OpenAI is winding down its fine-tuning platform, a key tool for developers seeking to customise models for specific tasks. The platform is no longer accessible to new users, effectively closing the door on fresh entrants to the service.
However, existing users of the fine-tuning platform will be able to continue creating training jobs for the coming months. While the exact duration of this wind-down period is not specified, the company has indicated that all fine-tuned models will remain available for inference until their base models are deprecated.
The changes were detailed in the company’s API documentation, which serves as the primary source for platform pricing and status updates. The move suggests a strategic consolidation of resources as OpenAI continues to iterate on its frontier models.
For institutions and developers relying on customised models, the transition period provides a window to manage dependencies before the platform is fully phased out. The conditional nature of the inference availability means that long-term planning will depend on the deprecation timeline of the underlying base models.
