A100
The A100 is a cloud GPU with 80 GB of memory and 300 W TDP. It is currently available from 7 providers starting at $0.27/hr, with a market median of $1.99/hr across 85 offerings.
Best deal
Hardware specifications
- Architecture
- Ampere
- VRAM
- 80 GB
- Memory bandwidth
- 1,935 GB/s
- FP16
- 312 TFLOPS
- FP32
- 20 TFLOPS
- TDP
- 300 W
Same across all providers.
Market comparison
13-day price history
Offerings
| Provider | Cloud | Count | vCPU | RAM | Region | Per GPU hr | SPOT | Total/hr | |
|---|---|---|---|---|---|---|---|---|---|
Alibaba Cloud GPU 0.0% 30dReferral link | — | ×4 | — | 378 GB | cn-shanghai | $0.73/hr | — | $2.91/hr | Launch ↗ |
HardwareHQ 0.0% 30dReferral link | — | ×4 | — | — | — | $1.19/hr | — | $4.76/hr | Launch ↗ |
Huawei Cloud GPU 0.0% 30dReferral link | — | ×4 | — | 256 GB | ap-southeast-2 | $4.40/hr | — | $17.59/hr | Launch ↗ |
Huawei Cloud GPU 0.0% 30dReferral link | — | ×4 | — | 256 GB | ap-southeast-2 | $4.40/hr | — | $17.59/hr | Launch ↗ |
Yandex Cloud 0.0% 30dReferral link | — | ×4 | — | — | RU (Moscow) | $4.45/hr | — | $17.80/hr | Launch ↗ |
LLMs that fit
| Model | Params | 16-bit | 8-bit | 4-bit |
|---|---|---|---|---|
DeepSeek R1 DeepSeek | 671B 37B active | Not released | Needs >8 GPUs | Barely fits on 8× |
DeepSeek V4 Flash DeepSeek | 284B | Not released | Not released | Barely fits on 3× |
Solar Open2 250B Upstage | 250B 15B active | Needs >8 GPUs | Barely fits on 5× | Barely fits on 3× |
Qwen3 235B-A22B Alibaba | 235B 22B active | Needs >8 GPUs | Barely fits on 5× | Barely fits on 3× |
Laguna-S 2.1 Poolside | 118B 8B active | 5× · 61.6 GB free | 3× · 54 GB free | 2× · 52 GB free |
gpt-oss-120b OpenAI | 117B 5.1B active | Not released | Not released | 2× · 66.4 GB free |
Llama 3.1 70B Meta | 70B | 3× · 72 GB free | 2× · 68.2 GB free | 1× · 26 GB free |
Qwen3.6 35B-A3B Alibaba | 35B 3B active | 2× · 56.8 GB free | 1× · 23.6 GB free | 1× · 46.4 GB free |
Qwen3 32B Alibaba | 32.8B | 2× · 65.2 GB free | 1× · 28.4 GB free | 1× · 50 GB free |
Qwen2.5 14B Alibaba | 14B | 1× · 44.6 GB free | 1× · 60.7 GB free | 1× · 68.6 GB free |
Llama 3.1 8B Meta | 8B | 1× · 59.1 GB free | 1× · 68.6 GB free | 1× · 73.3 GB free |
Mistral 7B Mistral AI | 7B | 1× · 61.8 GB free | 1× · 70 GB free | 1× · 74.1 GB free |
Estimates: model weights plus 20% for KV cache and runtime overhead, at short context lengths. Long contexts and large batches need more. Rows marked “tight” hold the weights but not that full margin. A ×N figure is the total VRAM across N of these GPUs and makes no claim about interconnect throughput.
Similar GPUs
See more GPUs like this in
Frequently asked questions
How much does the A100 cost per hour?
The cheapest A100 offering starts at $0.27/hr, and the market median is $1.99/hr.
Which cloud providers offer the A100?
We track A100 offerings from Alibaba Cloud GPU, HardwareHQ, Huawei Cloud GPU, Yandex Cloud.
How much VRAM does the A100 have?
The A100 has 80 GB of VRAM.
What LLMs can run on the A100?
With 80 GB of VRAM, the A100 can run 13 of our tracked models in 4-bit quantization, including DeepSeek R1, DeepSeek V4 Flash, Solar Open2 250B. 8 models fit in 16-bit precision.
What's the biggest LLM I can run on multiple A100s?
With 8× A100 (640 GB total), the largest tracked model that fits in 16-bit is Laguna-S 2.1 (118B params), which needs 5× GPUs.
Where is the A100 cheapest?
The cheapest A100 offering in our catalog is Alibaba Cloud GPU at $0.73/hr per GPU-hour (×4).