Together AI adds gemma-4-31B-it, GLM-5.1, and DeepSeek-V4-Pro to serverless

AI Impact Summary

Together AI has initiated a serverless model bring-up, adding google/gemma-4-31B-it, zai-org/GLM-5.1, deepSeek-V4-Pro, deepcogito/cogito-v2-1-671b, zai-org/GLM-4.5-Air-FP8, and zai-org/GLM-4.7 to their serverless offerings. This expansion of model availability is coupled with significant infrastructure updates, including the Slurm-on-Kubernetes cluster migration and improvements to worker daemon self-healing and job accounting. The shift to dynamic rate limits and prepaid billing represents a fundamental change in Together AI’s pricing model, requiring users to adjust their deployments and cost management strategies.