6GB / 12GB · AI Score 1.3/100 · first-party measured on 12 AI workloads
1.3 AI Score Includes estimates
Every number on this page is first-party: NVIDIA RTX A2000 was run on our pinned 12-workload AI suite on 2026-07-12, with under 0.5% run-to-run variance. 11 of the 12 workloads won't fit on 6GB / 12GB at the tested precision, Llama 3.1 8B, Qwen2.5-Coder 14B, Qwen3 32B, Llama 3.3 70B and others. We publish those as hard gates rather than quietly dropping to a smaller quant.
| Workload | Result | Telemetry | Data |
|---|---|---|---|
| Qwen3 4B | 68.47 tok/s | 2.8 GB peak57 W60°C1.19 tok/WQ4_K_M | ✓ Measured |
| Llama 3.1 8B | ✕ Won't fit needs ~8 GB | VRAM-gated at this precision | ✓ Measured |
| Nemotron Nano 9B v2 | ✕ Won't fit needs ~8 GB | VRAM-gated at this precision | Est. |
| Gemma 4 12B | ✕ Won't fit needs ~9 GB | VRAM-gated at this precision | Est. |
| Qwen2.5-Coder 14B | ✕ Won't fit needs ~11.5 GB | VRAM-gated at this precision | ✓ Measured |
| Qwen3 14B | ✕ Won't fit needs ~11 GB | VRAM-gated at this precision | Est. |
| Gemma 4 26B A4B | ✕ Won't fit needs ~20 GB | VRAM-gated at this precision | Est. |
| Qwen3 30B A3B | ✕ Won't fit needs ~20 GB | VRAM-gated at this precision | Est. |
| Gemma 4 31B | ✕ Won't fit needs ~22 GB | VRAM-gated at this precision | Est. |
| Qwen3 32B | ✕ Won't fit needs ~23 GB | VRAM-gated at this precision | ✓ Measured |
| Llama 3.3 70B | ✕ Won't fit needs ~46 GB | VRAM-gated at this precision | ✓ Measured |
| Workload | Result | Telemetry | Data |
|---|---|---|---|
| Sana 1.6B | ✕ Won't fit needs ~11 GB | VRAM-gated at this precision | Est. |
| Stable Diffusion XL | ✕ Won't fit needs ~11 GB | VRAM-gated at this precision | ✓ Measured |
| Playground v2.5 | ✕ Won't fit needs ~12 GB | VRAM-gated at this precision | Est. |
| Z-Image Turbo | ✕ Won't fit needs ~13 GB | VRAM-gated at this precision | ✓ Measured |
| PixArt-Sigma XL | ✕ Won't fit needs ~14 GB | VRAM-gated at this precision | Est. |
| FLUX.1 dev | ✕ Won't fit needs ~26 GB | VRAM-gated at this precision | ✓ Measured |
| Workload | Result | Telemetry | Data |
|---|---|---|---|
| FLUX.1 Kontext dev | ✕ Won't fit needs ~26 GB | VRAM-gated at this precision | ✓ Measured |
| Qwen-Image-Edit | ✕ Won't fit needs ~42 GB | VRAM-gated at this precision | ✓ Measured |
| Workload | Result | Telemetry | Data |
|---|---|---|---|
| LTX-Video (distilled) | ✕ Won't fit needs ~14 GB | VRAM-gated at this precision | ✓ Measured |
| Wan 2.2 5B (720p) | ✕ Won't fit needs ~18 GB | VRAM-gated at this precision | ✓ Measured |
| Architecture | Ampere |
| CUDA cores | 3,328 |
| VRAM | 6GB / 12GB GDDR6 |
| Memory bus | 192-bit |
| Memory bandwidth | 288 GB/s |
| Boost clock | 1,200 MHz |
| TDP | 70 W |
| Process | 8nm |
| Interface | PCIe 4.0 x8 |
| Release date | 2021-08-10 |
| Launch MSRP | $450 |
NVIDIA RTX A2000 scores 1.3/100, #102 of 102. It ran 1 of 12; 11 exceeded its 6GB / 12GB. Every figure here is our own measurement.
100% = this card, AI & Machine Learning headline metric (AI Score). #20 of 20 workstation cards in this vertical.
| GPU | Relative | % | AI Score |
|---|---|---|---|
| NVIDIA RTX 2000 Ada Generation | 238% | 3.1 | |
| AMD Radeon Pro W6800 | 231% | 3 | |
| AMD Radeon Pro W7800 | 231% | 3 | |
| Intel Arc Pro A60 | 138% | 1.8 | |
| NVIDIA RTX A2000 | 100% | 1.3 |
← All AI & Machine Learning GPU rankings
| Transistors | 12,000 million |
| Die size | 276 mm² |
| Process node | 8 nm |
| Fabricated by | Samsung |
| Transistor density | 43.5 million per mm² |
Denser than 78% of the 746 cards we have silicon data for. Density is the clearest measure of what a process node bought: a card that gained it without growing the die got its speed from the fab rather than the architecture.
Silicon figures from Wikipedia (CC BY-SA 4.0). Benchmarks on this page are our own. Compare every chip.
Whole-job timings, composed from our measured per-model results on this card.
Can't run: 3D game asset kit (needs Stable Diffusion XL), Product shoot, start to finish (needs Stable Diffusion XL), Product photo shoot (needs Stable Diffusion XL), Photo restoration batch (needs FLUX.1 Kontext dev), Restore and enlarge photos (needs FLUX.1 Kontext dev), Character sheet, 12 poses (needs FLUX.1 dev), Short social clips (needs Qwen3 32B), 60-second AI short film (needs Qwen3 32B), 24-frame storyboard (needs Qwen3 32B), 6-panel comic page (needs Qwen3 32B), Long-form article batch (needs Llama 3.3 70B), Full codebase review (needs Qwen2.5-Coder 14B).
This card is $450 to buy. The cheapest listed rate on Vast.ai is $0.110/hour, but that is the floor: we budget $0.132/hour, a 20% premium, because idle time, storage and unavailable cheap instances all land on the same bill. At that rate buying wins after 3,409 GPU-hours. Below it you are paying for idle silicon.
| How you would use it | GPU-hours a year | Rental cost a year | Time to break even |
|---|---|---|---|
| 2 hours a day, hobby | 730 | $96 | 4.7 years |
| 8 hours a day, working on it | 2,920 | $385 | 1.2 years |
| 24/7, always-on agent | 8,760 | $1,156 | 4.7 months |
At steady usage this card pays for itself inside a normal ownership window. Rental figures include a 20% premium over the cheapest listed rate. Ignores electricity, resale and the fact that a rented card can be a newer one tomorrow.
Cheapest of the RunPod and Vast on-demand rates we see, sampled daily. Spot and interruptible pricing runs lower.