DeepSeek-V4-Pro added to serverless — new model with 512K context

AI Impact Summary

DeepSeek-V4-Pro has been added to serverless, offering a new model option for users. This change introduces a context length of 512,000 tokens and a pricing structure of $2.10 per 1M input tokens, $4.40 per 1M output tokens, and $0.20 per 1M cached input tokens, utilizing FP4 quantization and supporting function calling and structured outputs. This addition expands Together AI’s serverless offerings and provides a new model for experimentation and deployment.