VESSL AI runs global inference infrastructure for Upstage's Solar Pro 4, supporting daily token growth from 8B to 155B (report)

Verified (fact)

VESSL AI said on Oct. 12 that it provided global inference infrastructure for Upstage's LLM Solar Pro 4 and applied its own optimization technology to improve processing speed and operating costs. Solar Pro 4 ranked third by actual usage on OpenRouter in September. VESSL AI said it supported the inference infrastructure needed as daily token throughput rose about 19-fold over that period, from 8 billion to 155 billion tokens.

Needs fact-check (claim)

VESSL AI claims processing speed improved by up to 3x. The comparison baseline and measurement conditions were not disclosed.

Why it matters (interpretation)

It is a case of a Korean neocloud handling the growing inference demand that comes when a Korean LLM is widely used overseas. In sovereign AI discussions, it shows that Korean companies can handle not just models but also serving infrastructure.

Sources

• Press 베슬AI, 업스테이지 '솔라 프로4' 글로벌 추론 뒷받침…처리속도 최대 3배↑ ZDNet Korea