Publicly listed prices across platforms as of September 28 (auto-refreshed hourly; full dataset available atReal-time price comparison homepage; prices are subject to the order page.
| GPU model | Today’s lowest monitored price | 24 hours | Platform | vs. yesterday | |
|---|---|---|---|---|---|
| RTX 3090 | ¥0.50/hour | ¥12/day | Vast.ai (Overseas) | — | Rent Now |
| T4 16GB | ¥1.08/hour | ¥26/day | Vast.ai (Overseas) | — | Rent Now |
| RTX 4090 | ¥1.24/hour | ¥30/day | Suanjiayun | — | Rent Now |
| RTX 5090 | ¥2.68/hour | ¥64/day | Suanjiayun | — | Rent Now |
| A6000 48GB | ¥3.53/hour | ¥85/day | Vast.ai (Overseas) | — | Rent Now |
| A100 80GB | ¥4.60/hour | ¥110/day | Vast.ai (Overseas) | — | Rent Now |
| H100 80GB | ¥9.60/hour | ¥230/day | Vast.ai (Overseas) | — | Rent Now |
Today’s Insights
· RTX 3090 daily rate starts at ¥12; Vast.ai (overseas) remains the top value choice for lightweight inference and fine-tuning 7B/13B models.
· RTX 4090 on Suanjiayun is listed at ¥30/day — more stable than Vast.ai’s spot pricing for the same GPU model, ideal for users requiring long-term stable training.
· T4 16GB daily rate is ¥26; Vast.ai (overseas) remains the go-to GPU for low-cost deployment of mid-size models like Llama-2-7B.
· A100 80GB at ¥110/day; H100 80GB up to ¥230/day. Vast.ai (overseas) consistently offers high-compute spot instances — large model pretraining users should monitor inventory closely.
· RTX 5090 launched at ¥64/day exclusively on Suanjiayun — ideal for experimental use cases demanding cutting-edge architecture with ample budget.