NVIDIA GeForce RTX 2080 Ti
Turing flagship. 11 GB GDDR6 at 616 GB/s — fits 13B Q4 comfortably, 7B Q4 at ~110-140 tok/s. Used $360-420 in 2026 makes it the 'enthusiast on a budget' floor; competes with used 3060 12 GB on raw VRAM and beats it on compute.
Affiliate disclosure: as an Amazon Associate and partner of other retailers, we earn from qualifying purchases. The verdict on this page is our editorial opinion; affiliate links never influence what we recommend.
Sub-scores sum to 518 / 1000. Headline = 518 × 0.70 (Estimated-confidence discount) = 363. This is an algorithmic performance-tier score — distinct from, and often lower than, the editorial “Our verdict” below, which weighs value and real-world fit (especially for hardware we haven’t measured yet). How scoring works →
Extrapolated from 616 GB/s bandwidth — 73.9 tok/s estimated. No measured benchmarks yet.
Plain-English: Best for 7B; 14B is tight — coding agent feels deliberate; vision models supported.
Verdicts extrapolated from catalog VRAM + bandwidth + ecosystem flags. Hover any chip for the rationale. Want measured numbers? Submit your own run with runlocalai-bench --submit.
This card is for the operator who needs 13B-class models at home and refuses to pay scalper prices for a 3090. The 2080 Ti runs 7B Q4 at 110-140 tok/s and 13B Q4 at 60-80 tok/s, both comfortably above interactive thresholds. 11 GB VRAM fits 13B Q4 with room for context, and 34B Q4 is possible with aggressive quantization (e.g., Q3_K_M) at 20-30 tok/s. What breaks: 11 GB is tight for 34B Q4 with long context, and Turing lacks official support for FP8 or newer attention kernels, so some inference frameworks may underperform. Pass on this card if you need 70B models or want native FP8 support; a used 3090 or 4070 Ti Super adds headroom. At $380 used, this is the cheapest entry to 13B-class local AI, but expect to tinker with quantization and context limits.
›Why this rating
The 2080 Ti offers exceptional value for 13B models at its used price, but its 11 GB VRAM and lack of modern features (FP8, sparse attention) limit it compared to newer cards. It's a strong budget enthusiast pick, not a top-tier workhorse.
Overview
Turing flagship. 11 GB GDDR6 at 616 GB/s — fits 13B Q4 comfortably, 7B Q4 at ~110-140 tok/s. Used $360-420 in 2026 makes it the 'enthusiast on a budget' floor; competes with used 3060 12 GB on raw VRAM and beats it on compute.
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Specs
| VRAM | 11 GB |
| Power draw (peak) | 250 W |
| Released | 2018 |
| MSRP | $1199 |
| Backends | CUDA Vulkan |
Models that fit
Open-weight models small enough to run on NVIDIA GeForce RTX 2080 Ti with usable context.
Frequently asked
What models can NVIDIA GeForce RTX 2080 Ti run?
Does NVIDIA GeForce RTX 2080 Ti support CUDA?
How much does NVIDIA GeForce RTX 2080 Ti cost?
Where next?
Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify hardware specifications.