NVIDIA GTX 1070 - 8GB usable
An entry into 8GB CUDA, adequate for 7B models and learning the stack. Pascal predates tensor cores, so prompts process slowly.
Specifications
| Brand | NVIDIA |
|---|---|
| Model | GTX 1070 |
| Usable VRAM | 8GB |
| Architecture | Pascal |
| CUDA / Stream Processors | 1,920 |
| Memory Bandwidth | 256 GB/s |
| TDP | 150W |
| FP32 TFLOPS | 6.5 |
Current Offers
Prices last updated:
GPUDojo is reader-supported. When you buy through links on our site, we may earn an affiliate commission.
Lowest tracked offers by retailer
UK · GBP · Weekly since Aug 2026 · Shipping may be extra
- eBay £84 at 90-day high
About these prices
Each line is the lowest offer we tracked from that retailer at weekly snapshots, ending at today's offer. The condition can vary between snapshots (new, used, or unrecorded for older ones). eBay prices exclude shipping.
Weekly offers table
| Date | eBay |
|---|---|
| Today, Nov 3, 2026 | £83.90 |
| Oct 30, 2026 | £79.99 |
| Oct 23, 2026 | £70.00 |
| Oct 9, 2026 | £75.00 |
| Oct 2, 2026 | £70.00 |
| Sep 25, 2026 | £73.50 |
| Sep 18, 2026 | £73.50 |
| Sep 11, 2026 | £75.00 |
| Sep 4, 2026 | £70.00 |
| Sep 1, 2026 | £66.22 |
| Aug 28, 2026 | £63.10 |
| Aug 21, 2026 | £57.90 |
For AI / LLM Use
Limited VRAM restricts you to 7B quantized models. Slower generation, usable but not snappy. Older architecture may have limited software support (check CUDA compatibility).
What Models Can It Run?
- 7B Q6_K, 14B Q3_K (tight)
- 7B Q4_K_M only
Estimated Performance
Generation: ~19 tokens/sec
Prefill: ~116 tokens/sec
Recommended Quantisations
- Q4_K_M for 7B models
- Q3_K for larger experiments
Pros & Cons
Pros
- Only 150W TDP: power efficient
- Consumer card: easy to install, display output
Cons
- Only 8GB usable VRAM: limited to small models
- Low memory bandwidth: slower token generation
- Older Pascal architecture: verify current CUDA support
Community Verdict
No community reviews yet for the GTX 1070. Know a good review? Let us know.