Together AI launches Provisioned Throughput for open models

Provisioned Throughput offers reserved inference capacity with token-based pricing and a 99% uptime SLA, available today for MiniMax M3 and GLM-5.2. Together AI claims up to 90% lower cost than Claude Opus 4.8, with a one-month minimum term. The company's monthly token volume has grown from 30B to 400T in nine months.