Expired50% confidenceFactExact time
Domestic platforms such as AutoDL and Hengyuan Cloud charge by the hour, with the cost of using an A100 as low as a few yuan RMB per hour.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/11/2026
First Seen
Valid until: 9/25/2026(expired)
Sources
Related Entities
Related Claims
VerifiedRunning inference services on a 70B-parameter open-source model requires at least 2–4 A100/H100-class GPUs, with cloud rental costs of $5–20 per hour.73% similarUnverifiedUsing GPU cloud providers such as Lambda Cloud and CoreWeave, the monthly rental cost for a single A100 can be as low as $1,000–$1,500.70% similarUnverifiedTaking a 2TB audio library (approximately 35,000 hours) as an example, using cloud-based ASR services at roughly $1–2 per hour, the total cost could reach tens of thousands of dollars.70% similarUnverifiedCloud platform spot instance prices are typically only 10%–30% of on-demand instance prices; for example, on AWS an A10G GPU on-demand price is approximately $1.50 per hour, while the spot price can be as low as $0.30–$0.50.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/896980API
curl https://kongchang.com/api/v1/knowledge/claims/896980MCP
get_claim(id=896980)