B300
The B300 is a cloud GPU with 288 GB of memory and 1400 W TDP. It is currently available from 6 providers starting at $6.94/hr, with a market median of $13.88/hr across 13 offerings.
Best deal
Hardware specifications
- Architecture
- Blackwell
- VRAM
- 288 GB
- Memory type
- HBM3
- Memory bandwidth
- 8,000 GB/s
- FP16
- 2500 TFLOPS
- FP32
- 75 TFLOPS
- TDP
- 1400 W
Same across all providers.
Market comparison
9-day price history
Offerings
LLMs that fit
| Model | Params | 16-bit | 8-bit | 4-bit |
|---|---|---|---|---|
Kimi K3 Moonshot AI | 2.8T 104B active | Not released | Not released | Barely fits on 8× |
DeepSeek V4 Pro DeepSeek | 1.6T 49B active | Not released | Not released | 5× · 13.5% spare |
Kimi K2.5 Moonshot AI | 1T 32B active | Not released | Not released | Barely fits on 3× |
GLM-5.2 Z.ai | 753B | Barely fits on 8× | 5× · 247.2 GB free | 3× · 169.2 GB free |
DeepSeek R1 DeepSeek | 671B 37B active | Not released | 4× · 14% spare | 3× · 286.8 GB free |
DeepSeek V4 Flash DeepSeek | 284B | Not released | Not released | 1× · 57.6 GB free |
Solar Open2 250B Upstage | 250B 15B active | 3× · 142.8 GB free | 2× · 180 GB free | 1× · 57.6 GB free |
Qwen3 235B-A22B Alibaba | 235B 22B active | 3× · 187.2 GB free | 2× · 204 GB free | 1× · 70.8 GB free |
Laguna-S 2.1 Poolside | 118B 8B active | 2× · 237.6 GB free | 1× · 102 GB free | 1× · 180 GB free |
gpt-oss-120b OpenAI | 117B 5.1B active | Not released | Not released | 1× · 194.4 GB free |
Llama 3.1 70B Meta | 70B | 1× · 120 GB free | 1× · 196.2 GB free | 1× · 234 GB free |
Qwen3.6 35B-A3B Alibaba | 35B 3B active | 1× · 184.8 GB free | 1× · 231.6 GB free | 1× · 254.4 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 B300 cost per hour?
The cheapest B300 offering starts at $6.94/hr, and the market median is $13.88/hr.
Which cloud providers offer the B300?
We track B300 offerings from HardwareHQ, Vast.ai.
How much VRAM does the B300 have?
The B300 has 288 GB of VRAM.
What LLMs can run on the B300?
With 288 GB of VRAM, the B300 can run 17 of our tracked models in 4-bit quantization, including Kimi K3, DeepSeek V4 Pro, Kimi K2.5. 11 models fit in 16-bit precision.
What's the biggest LLM I can run on multiple B300s?
With 8× B300 (2304 GB total), the largest tracked model that fits in 16-bit is Solar Open2 250B (250B params), which needs 3× GPUs.
Where is the B300 cheapest?
The cheapest B300 offering in our catalog is HardwareHQ at $6.94/hr per GPU-hour (×2).