The 5060 Ti at £119/mo and the 6000 Pro at £899/mo are 6.5× the cost. Most teams should not jump directly between them — the 5090 at £399/mo is the natural intermediate step. The 6000 Pro is the right answer only for specific 70B-class workloads.
TL;DR
Skip the 5060 Ti → 6000 Pro upgrade unless you genuinely need 70B FP8 single-card, ECC for compliance, or 96 GB to stack many models. For most upgrade paths, go 5060 Ti → 5090 (32 GB) first.
Capability delta
| Spec | RTX 5060 Ti | RTX 6000 Pro | Delta |
|---|---|---|---|
| VRAM | 16 GB | 96 GB ECC | +500% |
| Memory bandwidth | 448 GB/s | 1,792 GB/s | +300% |
| FP16 TFLOPS | ~24 | ~234 | +875% |
| Monthly cost | £119 | £899 | +550% |
Upgrade triggers
Genuine reasons to skip 5090 and go straight to 6000 Pro:
- You need to host Llama 3.3 70B FP8 single-card (5090 can’t)
- Compliance requires ECC + certified drivers
- You want to stack 5+ models on one card
- You need full SFT of 13B+ models (5090 too tight)
Why 5090 is usually the better intermediate step
For 95% of upgrade paths from 5060 Ti, the 5090 is the right next step:
- 2× VRAM (16 → 32 GB) covers most growth
- 3× the throughput at 2× the cost — better marginal economics
- 32 GB fits 14B FP16 and 70B INT4
- £399/mo vs £899/mo
Bottom line
Upgrade in steps: 5060 Ti → 5090 → 6000 Pro. Skip the middle only when you have a specific 70B-class or compliance need.
