Together AI has launched Provisioned Throughput, a new service offering reserved inference capacity for open-weight AI models. This service provides token-based pricing with a 99% uptime Service Level Agreement (SLA), aiming to offer the predictability of proprietary models with the cost-effectiveness of open-source alternatives. Initially available for MiniMax M3 and GLM-5.2, Provisioned Throughput is designed to significantly reduce inference costs, potentially by up to 90% compared to models like Claude Opus 4.8. AI
IMPACT Offers a more predictable and cost-effective way for businesses to utilize open-weight AI models for production workloads.
RANK_REASON This is a new product offering from an AI infrastructure provider, not a frontier model release or significant industry-wide event.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
