Together AI expands its fine-tuning platform and cuts training prices, competing directly with OpenAI and proprietary-model vendors.

Together AI announced via LinkedIn that it is expanding its Together Fine-Tuning platform to support additional open-weight models, including GLM-5.3 and Kimi K2.7-Code. The update also notes enhancements to the training loop designed to make model development more transparent and iterative for users.

New capabilities highlighted in the announcement include live run tracking, experiment comparison, early stopping based on validation loss, and pre-training inspection of tokenized data. These tools underscore a focus on improving developer efficiency and control, potentially making the platform more appealing to enterprise and research clients.

Alongside these feature upgrades, Together AI has cut training prices by 30–70% on selected models. While this reduction lowers adoption barriers for cost-sensitive users and could stimulate higher usage volumes, it may also compress margins, requiring greater scale efficiencies to maintain profitability.

The update positions Together AI more competitively within the AI infrastructure and fine-tuning market, particularly against rivals offering both open and proprietary model tooling. For investors, the combination of expanded model support, enhanced tooling, and aggressive pricing signals a clear strategy to capture market share and deepen customer engagement across the rapidly growing AI development sector.