NVIDIA RTX 3070 Ti GPU: Specs, VRAM, Price & Benchmarks (2026)
The RTX 3070 Ti is a real GPU (launched 2021), but no provider on Aquanode is listing it for rental right now, so there is no live hourly price to quote. Availability changes as providers add and retire hardware; the specs, VRAM and price context below still apply.
How much VRAM does the RTX 3070 Ti have?
The RTX 3070 Ti has 8GB GDDR6X, with 608 GB/s of peak memory bandwidth.
RTX 3070 Ti VRAM calculator: check which models fit in its memory at each precision.
All models that fit in 8 GB: the open models whose weights and overhead fit, at native, FP8 and INT4 precision.
What can the RTX 3070 Ti run?
Popular open models from small to frontier scale, with the memory each needs and how many RTX 3070 Ti cards (8GB GDDR6X each) that takes.
| Model | As published | FP8 | INT4 |
|---|---|---|---|
| Qwen/Qwen3-8B 8.2B | BF16: ~18.3 GB, 3 GPUs | FP8: not supported | INT4: ~4.6 GB, 1 GPU |
| Qwen/Qwen2.5-14B-Instruct 14.8B | BF16: ~33 GB, 5 GPUs | FP8: not supported | INT4: ~8.3 GB, 2 GPUs |
| Qwen/Qwen3-32B 32.8B | BF16: ~73.2 GB, 10 GPUs | FP8: not supported | INT4: ~18.3 GB, 3 GPUs |
| Qwen/Qwen-72B 72.3B | BF16: ~162 GB, 21 GPUs | FP8: not supported | INT4: ~40.4 GB, 6 GPUs |
| MiniMaxAI/MiniMax-M2.7 228.7B | FP8: not supported | – | INT4: ~128 GB, 16 GPUs |
| deepseek-ai/DeepSeek-R1 684.5B | FP8: not supported | – | INT4: ~383 GB, 48 GPUs |
Estimates: weights at the stated precision plus a flat 20% for KV cache and overhead, at a moderate context length. A dash means the precision is not offered for that model (it is already published at that size). INT4 needs a published quantized checkpoint. Open any model for a per-GPU breakdown, or use the RTX 3070 Ti VRAM calculator.
RTX 3070 Ti specs
| Architecture | launched 2021 |
| VRAM | 8GB GDDR6X |
| Memory bandwidth | 608 GB/s |
| FP16 / BF16 tensor throughput | 87 TFLOPS (peak, dense) |
| TDP | 290W |
| Form factor | PCIe |
Specs sourced from the vendor's public product page. See the source.
GPU Glossary: What is VRAM?, Tensor Cores, CUDA Cores, TFLOPS
Good for
608 GB/s of memory bandwidth on 8GB makes small quantized models fast: a 7-8B model at INT4 fits with context to spare and decodes quickly for the card's size.
Not good for
8GB is the whole story: a 7B model at FP16 needs about 14GB and does not fit, and nothing above ~13B fits even at INT4. No FP8 tensor cores, no NVLink.
Related guides
Other models in the same generation. The full list is in the GPU index.
Get notified when the price drops
GPU supply moves hourly. Tell us what you're waiting for and we'll email you when a matching offer appears across any provider we track.
