Unverified50% confidenceBenchmarkExact time
In the AgentX test, vLLM achieved revenue of $38.96 per GPU hour.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/29/2026
First Seen
Valid until: 11/12/2026
Sources
Related Entities
Related Claims
Unverified训练DeepSeek-V3仅消耗约278万H800 GPU小时，成本约合550万美元66% similarUnverifiedAt the prices the author actually paid ($0.414 per hour for the 3090, $0.826 per hour for the L40S) and counting only GPU busy time, at low batch sizes the 3090's cost per million output tokens is about half that of the L40S.65% similarUnverifiedBeyond the break-even volume, the actual cost of a 30B-class model on a busy GPU is approximately $0.06–$0.85 per million tokens.65% similarVerifiedAccording to DeepSeek's data, each decode GPU produces about 14,000 tokens per second for all of its users, bringing in about $38 of revenue per GPU per hour on average.64% similarUnverifiedIf a team's sustained throughput exceeds about 40 to 50 million tokens per day, renting dedicated vLLM/SGLang GPU instances is usually another 45%–60% cheaper than token-billed serverless pricing.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1110154API
curl https://kongchang.com/api/v1/knowledge/claims/1110154MCP
get_claim(id=1110154)