Cerebras and Upstage have collaborated to bring ultra-fast, production-ready AI to users, with Upstage’s Solar 31B model running on the Cerebras Wafer-Scale Engine achieving inference speeds of up to 2,000 tokens per second. Upstage, a leading Korean AI company, offers enterprise AI products across large language models, document intelligence, and AI workflow automation. Sung Kim, CEO of Upstage, stated, “We’re excited to work with Cerebras to bring our Solar models to developers with industry-leading inference speeds.” Julie Choi, Chief Marketing Officer of Cerebras, said, “South Korea is at the forefront of AI innovation, and we’re thrilled to partner with Upstage.” The partnership targets mission-critical enterprise workflows in South Korea’s finance, insurance, healthcare, and manufacturing industries. Cerebras Inference provides OpenAI-compatible APIs, allowing developers to build with minimal code changes. An example showed Solar 31b taking 246 sources of information and executing a deep research query, while a comparable query with Sonnet 4.6 with a similar number of sources takes eight minutes.