Together AI adds Qwen/Qwen3-Coder-Next-FP8 and Google Gemma models to serverless

AI Impact Summary

Together AI has expanded its serverless model offerings with the addition of Qwen/Qwen3-Coder-Next-FP8, DeepSeek-V4-Pro, and several Google Gemma models including gemma-4-31B-it and gemma-3n-E4B-it. This expansion, coupled with the deprecation of older models like Qwen3-VL-8B-Instruct and others, signals a shift in Together AI’s model portfolio towards more recent and performant offerings. The transition to prepaid billing and dynamic rate limits further streamlines the cost management for users.