datacenter_gpu
NVIDIA A800 80GB
Facts
- VRAM
- 80 GB 1 Source
- Reference rental (CNY/h)
- ¥5.88 /h ($0.87)Live · AutoDL pay-as-you-go median (China) · 2026-10-19 · Sinan Compute ↗
- All public quotes
- autodl on-demand min $0.87 (93 hosts) · autodl on-demand median $0.87 (93 hosts)
- Rental source
- Unverified
- Note
- China-market version of the A100 80GB (reduced NVLink bandwidth); rentable pay-as-you-go on Chinese compute platforms
cost = p50 ms ÷ 1000 × rental ÷ 3600 × 1000
Sources · 1 sources
name,vram_gb https://www.nvidia.cn/data-center/a800/ Official page · Fetched 2026-09-18 · No snapshot stored yet NVIDIA China official site A800 product page
Latency ranking on this card
| Model | p50 ms | Cost per 1k inferences ¥ | samples |
|---|---|---|---|
| NORA Declare Lab (SUTD) | 146 | 0.238 | 1 |
| OpenVLA Stanford / UC Berkeley / TRI et al. (OpenVLA team) | 187 | 0.306 | 1 |
| SmolVLA Hugging Face (LeRobot) | 226 | 0.368 | 1 |
| SpatialVLA Shanghai AI Laboratory et al. | 451 | 0.736 | 1 |
| MolmoAct Allen Institute for AI（Ai2） | 937 | 1.53 | 1 |
| Isaac GR00T N1.7 NVIDIA | — | — no latency data | No data |
| GR00T N1.6 3B NVIDIA | — | — no latency data | No data |
| LingBot-VLA 2.0 Robbyant (Ant Group) | — | — no latency data | No data |
| GO-1 (Genie Operator-1) AgiBot | — | — no latency data | No data |
| π0 (pi-zero) Physical Intelligence | — | — no latency data | No data |
| π0.5 Physical Intelligence | — | — no latency data | No data |
| X-VLA Tsinghua University AIR (2toINF) | — | — no latency data | No data |
| Octo UC Berkeley RAIL et al. | — | — no latency data | No data |
| RDT-1B Tsinghua University TSAIL (thu-ml) | — | — no latency data | No data |
| CogACT Microsoft Research | — | — no latency data | No data |
| UniVLA OpenDriveLab (HKU et al.) | — | — no latency data | No data |
| RynnVLA-001 Alibaba DAMO Academy | — | — no latency data | No data |
| WALL-OSS X Square Robot | — | — no latency data | No data |
| InternVLA-M1 Shanghai AI Laboratory InternRobotics | — | — no latency data | No data |
| GraspVLA Peking University EPIC Lab / Galbot | — | — no latency data | No data |
| Being-H0 BeingBeyond (Beijing) | — | — no latency data | No data |
| EO-1 IPEC-COMMUNITY (Shanghai AI Laboratory affiliated) | — | — no latency data | No data |
| LingBot-VLA 4B Robbyant (Ant Group) | — | — no latency data | No data |
| VLA-Adapter OpenHelix team | — | — no latency data | No data |
| HPT（Heterogeneous Pre-trained Transformers） MIT (Lirui Wang, Kaiming He et al.) | — | — no latency data | No data |
| LAPA KAIST / University of Washington et al. (Latent Action Pretraining) | — | — no latency data | No data |
| LLaVA-VLA OpenHelix team | — | — no latency data | No data |
| RoboVLMs（KosMos P.H.） Tsinghua University / ByteDance Research / CASIA et al. | — | — no latency data | No data |
| Magma-8B Microsoft Research | — | — no latency data | No data |
| V-JEPA 2-AC Meta FAIR | — | — no latency data | No data |
| DexVLA Midea AI Research et al. (DexVLA team) | — | — no latency data | No data |
| GR-1 ByteDance Research | — | — no latency data | No data |
| Spirit v1.5 Spirit AI | — | — no latency data | No data |
| Galaxea G0.5 Galaxea | — | — no latency data | No data |
| RDT2-VQ Tsinghua University TSAIL (thu-ml) | — | — no latency data | No data |
| InternVLA-A1.5 Shanghai AI Laboratory InternRobotics | — | — no latency data | No data |
| VLA-0 NVIDIA Research | — | — no latency data | No data |
| MiniVLA Stanford ILIAD | — | — no latency data | No data |
| Genie Envisioner（GE-Act） AgiBot | — | — no latency data | No data |
| GR00T N1.5 3B NVIDIA | — | — no latency data | No data |
Clear direction, grounded judgement.