NVIDIA RTX 3090 GPU: Specs, VRAM, Price & Benchmarks (2026)
The NVIDIA RTX 3090 (24GB GDDR6X) launched in 2020. This guide covers its full specs, VRAM, SXM/PCIe differences where they apply, AI performance, and live per-GPU rental pricing on Aquanode.
Short answer on cost: the RTX 3090 rents from $0.150 per GPU per hour on Aquanode.
How much VRAM does the RTX 3090 have?
The RTX 3090 has 24GB GDDR6X, with 936 GB/s of peak memory bandwidth.
- VRAM: RTX 3090 24GB GDDR6X vs RTX 4090 24GB GDDR6X
- VRAM: RTX 3090 24GB GDDR6X vs RTX 3080 10GB GDDR6X (a 12GB variant also ships)
What fits in 24GB GDDR6X of VRAM
| Model | Precision | Fits? |
|---|---|---|
| Llama 3 8B | BF16 | roughly 16GB. Fits with a usable KV cache. |
| Mixtral 8x7B | INT4 | ~24GB. Right at the limit; expect to need offload. |
| Llama 3.1 70B | INT4 | ~35-40GB. Does not fit on one card; fits across two. |
| Llama 3.1 70B | BF16 | ~140GB. Needs 6 cards. |
Approximate, based on published parameter counts and standard bytes-per-parameter rules of thumb (FP16 ≈ 2 bytes/param, INT4 ≈ 0.5-0.6 bytes/param). Real footprint also depends on KV-cache size and framework overhead.
What can the RTX 3090 run?
Popular open models from small to frontier scale, with the memory each needs and how many RTX 3090 cards (24GB GDDR6X each) that takes.
| Model | As published | FP8 | INT4 |
|---|---|---|---|
| Qwen/Qwen3-8B 8.2B | BF16: ~18.3 GB, 1 GPU | FP8: not supported | INT4: ~4.6 GB, 1 GPU |
| Qwen/Qwen2.5-14B-Instruct 14.8B | BF16: ~33 GB, 2 GPUs | FP8: not supported | INT4: ~8.3 GB, 1 GPU |
| Qwen/Qwen3-32B 32.8B | BF16: ~73.2 GB, 4 GPUs | FP8: not supported | INT4: ~18.3 GB, 1 GPU |
| Qwen/Qwen-72B 72.3B | BF16: ~162 GB, 7 GPUs | FP8: not supported | INT4: ~40.4 GB, 2 GPUs |
| MiniMaxAI/MiniMax-M2.7 228.7B | FP8: not supported | – | INT4: ~128 GB, 6 GPUs |
| deepseek-ai/DeepSeek-R1 684.5B | FP8: not supported | – | INT4: ~383 GB, 16 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 3090 VRAM calculator.
RTX 3090 VRAM calculator: check which models fit in its memory at each precision.
All models that fit in 24 GB: the open models whose weights and overhead fit, at native, FP8 and INT4 precision.
RTX 3090 specs
| Architecture | NVIDIA, launched 2020 |
| VRAM | 24GB GDDR6X |
| Memory bandwidth | 936 GB/s |
| FP16 / BF16 tensor throughput | 142 TFLOPS (peak, dense) |
| Interconnect | NVLink, 2-way |
| TDP | 350W |
| Form factor | PCIe, 3-slot |
Specs sourced from the vendor's public datasheet/product page. See the source.
Related reading: RTX 3090 for AI.
GPU Glossary: What is VRAM?, Tensor Cores, CUDA Cores, TFLOPS, NVLink vs PCIe
RTX 3090 AI performance
24GB and 936 GB/s at the cheapest hourly rate of any 24GB card here, plus a working NVLink bridge. Two of these give 48GB with a real GPU-to-GPU path, which no other consumer card on this marketplace can do.
- Dense FP16/BF16 tensor throughput: 142 TFLOPS
- Memory bandwidth: 936 GB/s
Ampere: no FP8 tensor cores at all, so FP8 serving is off the table and it does roughly a quarter of an RTX 4090's effective inference throughput on quantized workloads. No ECC, 350W, and a 3-slot card.
The cheapest RTX 3090 offer right now ($0.150/GPU/hr) is about 48% below the market median of $0.286/GPU/hr.
RTX 3090 price: what does it cost?
Buying. $1,499 (launch MSRP).
Renting. The cheapest current on-demand rate for the RTX 3090 on Aquanode is $0.150/GPU/hr (no on-demand supply right now; this is the cheapest offer overall). Live rates range from $0.150 to $0.676 per GPU per hour, with a median of $0.286/GPU/hr. Billed by the offer's own terms; the table below shows every live rate.
| Region | $/GPU/hr | Available | VRAM | vCPU | RAM |
|---|---|---|---|---|---|
| Sfo, Us | $0.150 | 1 | 24 GB | 102 | 217 GB |
| United States | $0.220 | 1 | 24 GB | 32 | 60 GB |
| Jiangsu, Cn | $0.276 | 4 | 24 GB | 24 | 31.5 GB |
RTX 3090 price history
Aquanode stores one snapshot of its GPU price index per UTC day. For the RTX 3090 that is 25 days so far, 2026-10-06 to 2026-10-30, so this is a short history, not a long-run trend. The lowest per-GPU rate went from $0.198 on 2026-10-06 to $0.220 on 2026-10-30.
| Day (UTC) | Lowest per GPU hour | Median per GPU hour | Data-center lowest per GPU hour | Offers |
|---|---|---|---|---|
| 2026-10-30 | $0.220 | $0.519 | – | 14 |
| 2026-10-29 | $0.150 | $0.486 | – | 11 |
| 2026-10-28 | $0.150 | $0.268 | – | 6 |
| 2026-10-27 | $0.353 | $0.549 | – | 10 |
| 2026-10-26 | $0.360 | $0.502 | – | 6 |
| 2026-10-25 | $0.198 | $0.220 | – | 17 |
| 2026-10-24 | $0.150 | $0.500 | – | 6 |
| 2026-10-23 | $0.346 | $0.641 | – | 3 |
| 2026-10-22 | $0.626 | $0.685 | – | 3 |
| 2026-10-21 | $0.150 | $0.626 | – | 3 |
| 2026-10-20 | $0.150 | $0.685 | – | 5 |
| 2026-10-19 | $0.162 | $0.615 | – | 10 |
| 2026-10-18 | $0.162 | $0.562 | – | 15 |
| 2026-10-17 | $0.198 | $0.451 | – | 12 |
| 2026-10-16 | $0.242 | $0.297 | $0.550 | 22 |
| 2026-10-15 | $0.162 | $0.297 | $0.550 | 24 |
| 2026-10-14 | $0.198 | $0.297 | $0.550 | 25 |
| 2026-10-13 | $0.198 | $0.297 | $0.550 | 28 |
| 2026-10-12 | $0.185 | $0.297 | $0.550 | 34 |
| 2026-10-11 | $0.185 | $0.242 | $0.550 | 47 |
| 2026-10-10 | $0.198 | $0.297 | $0.550 | 28 |
| 2026-10-09 | $0.200 | $0.242 | $0.550 | 41 |
| 2026-10-08 | $0.198 | $0.242 | $0.550 | 39 |
| 2026-10-07 | $0.198 | $0.242 | $0.550 | 43 |
| 2026-10-06 | $0.198 | $0.297 | $0.550 | 30 |
Each row is the stored daily snapshot of the live index, copied as recorded. A dash means that day stored no figure. The same series is available as JSON and summarised in the monthly GPU price report.
How this price is calculated
All prices on this page are normalized to a per-GPU hourly rate using each offer's authoritative GPU count, so that raw price is divided by the number of GPUs it actually covers; some offers report price as already per-GPU, so those are used as-listed. An offer with a missing, zero, or invalid GPU count is excluded entirely rather than published at a guessed rate.
No offers were excluded from this snapshot for a missing or invalid price. 1 offer(s) were dropped as high-side price outliers (more than 3x the model's median rate).
Only the cheapest qualifying offer per provider is shown in the table above. This page regenerates at most once per hour.
RTX 3090 vs RTX 4090: how do they compare?
- VRAM: RTX 3090 24GB GDDR6X vs RTX 4090 24GB GDDR6X
- Memory bandwidth: 936 GB/s vs 1,008 GB/s
On Aquanode right now, RTX 3090 starts at $0.150/GPU/hr against RTX 4090's $0.393/GPU/hr, about 62% less.
RTX 3090 vs RTX 3080: how do they compare?
- VRAM: RTX 3090 24GB GDDR6X vs RTX 3080 10GB GDDR6X (a 12GB variant also ships)
- Memory bandwidth: 936 GB/s vs 760 GB/s
- Dense FP16 tensor throughput: 142 TFLOPS vs 119 TFLOPS
On Aquanode right now, RTX 3090 starts at $0.150/GPU/hr against RTX 3080's $0.176/GPU/hr, about 15% less.
Compare the RTX 3090 with other GPUs
Compare RTX 3090 with
- B200 vs RTX 3090
- H200 vs RTX 3090
- H100 vs RTX 3090
- A100 vs RTX 3090
- DGX A100 vs RTX 3090
- RTX 3090 vs V100
- AMD MI300X vs RTX 3090
- L40S vs RTX 3090
- L40 vs RTX 3090
- A40 vs RTX 3090
- RTX 3090 vs RTX A6000
- RTX 3090 vs RTX A5000
- RTX 3090 vs RTX A4000
- L4 vs RTX 3090
- RTX 3090 vs T4
- RTX 3090 vs RTX 6000 Ada
- RTX 3090 vs RTX PRO 6000
- RTX 3090 vs RTX PRO 6000 WS
- RTX 3090 vs RTX PRO 6000 SE
- RTX 3090 vs RTX PRO 5000
- RTX 3090 vs RTX 5090
- RTX 3090 vs RTX 5080
- RTX 3090 vs RTX 5070 Ti
- RTX 3090 vs RTX 5070
- RTX 3090 vs RTX 5060 Ti
- RTX 3090 vs RTX 4090
- RTX 3090 vs RTX 4080 Super
- RTX 3090 vs RTX 4080
- RTX 3090 vs RTX 4070 Ti
- RTX 3090 vs RTX 4070 Super
- RTX 3090 vs RTX 4070
- RTX 3090 vs RTX 4060 Ti
- RTX 3080 vs RTX 3090
- RTX 3070 vs RTX 3090
- RTX 3060 vs RTX 3090
- B300 vs RTX 3090
- AMD MI355X vs RTX 3090
- AMD MI325X vs RTX 3090
- H100 NVL vs RTX 3090
- RTX 3090 vs RTX PRO 4500
- RTX 3090 vs RTX PRO 4500 SE
- RTX 3090 vs RTX PRO 4000
- RTX 3090 vs RTX 5880 Ada
- RTX 3090 vs RTX 5000 Ada
- RTX 3090 vs RTX 4000 Ada
- RTX 3090 vs RTX 4000 SFF Ada
- RTX 2000 Ada vs RTX 3090
- RTX 3090 vs RTX 4070 Ti Super
- RTX 3090 vs RTX A4500
- RTX 2080 Ti vs RTX 3090
- RTX 3080 Ti vs RTX 3090
- RTX 3070 Ti vs RTX 3090
- RTX 3060 Ti vs RTX 3090
- GTX 1080 Ti vs RTX 3090
- RTX 3090 vs RTX 4060
- RTX 3090 vs RTX 5060
- GTX 1660 Super vs RTX 3090
- RTX 2060 vs RTX 3090
- Radeon RX 9070 XT vs RTX 3090
- Radeon RX 6700 XT vs RTX 3090
- RTX 3090 vs RTX 6000
- A16 vs RTX 3090
- P40 vs RTX 3090
- P4 vs RTX 3090
- Quadro P2000 vs RTX 3090
- Quadro M4000 vs RTX 3090
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.
Rent a RTX 3090 on Aquanode
- On-demand instances from $0.150/GPU/hr, billed by the provider's own terms, with no hardware procurement or long-term commitment.
- 28 live offers across 3 regions today.
- Set a price/availability alert above to hear the moment a cheaper or newly-available RTX 3090 offer appears.
- Compare every RTX 3090 offer side by side, or browse the full multi-provider GPU marketplace.
Good for
24GB and 936 GB/s at the cheapest hourly rate of any 24GB card here, plus a working NVLink bridge. Two of these give 48GB with a real GPU-to-GPU path, which no other consumer card on this marketplace can do.
Not good for
Ampere: no FP8 tensor cores at all, so FP8 serving is off the table and it does roughly a quarter of an RTX 4090's effective inference throughput on quantized workloads. No ECC, 350W, and a 3-slot card.
RTX 3090 FAQs
How much VRAM does the RTX 3090 have?
The RTX 3090 has 24GB GDDR6X, with 936 GB/s of peak memory bandwidth.
What is the NVIDIA RTX 3090?
The NVIDIA RTX 3090 is a GPU released in 2020, with 24GB GDDR6X of memory and a 350W power envelope. See the full spec table above for interconnect, form factor and tensor-throughput details.
How much does it cost to rent a RTX 3090?
Live RTX 3090 rental prices currently range from $0.150 to $0.676 per GPU per hour, with a median of $0.286 per GPU per hour.
What's the cheapest RTX 3090 rate?
The lowest current RTX 3090 rate on Aquanode is $0.150 per GPU per hour in Sfo, Us.
How does the RTX 3090 compare to the RTX 4090?
VRAM: RTX 3090 24GB GDDR6X vs RTX 4090 24GB GDDR6X Memory bandwidth: 936 GB/s vs 1,008 GB/s On Aquanode right now, RTX 3090 starts at $0.150/GPU/hr against RTX 4090's $0.393/GPU/hr, about 62% less. See the full RTX 3090 vs RTX 4090 comparison for a shared-provider price breakdown.
How does the RTX 3090 compare to the RTX 3080?
VRAM: RTX 3090 24GB GDDR6X vs RTX 3080 10GB GDDR6X (a 12GB variant also ships) Memory bandwidth: 936 GB/s vs 760 GB/s Dense FP16 tensor throughput: 142 TFLOPS vs 119 TFLOPS On Aquanode right now, RTX 3090 starts at $0.150/GPU/hr against RTX 3080's $0.176/GPU/hr, about 15% less. See the full RTX 3090 vs RTX 3080 comparison for a shared-provider price breakdown.
What is the RTX 3090 good for?
24GB and 936 GB/s at the cheapest hourly rate of any 24GB card here, plus a working NVLink bridge. Two of these give 48GB with a real GPU-to-GPU path, which no other consumer card on this marketplace can do.
What are the RTX 3090's limitations?
Ampere: no FP8 tensor cores at all, so FP8 serving is off the table and it does roughly a quarter of an RTX 4090's effective inference throughput on quantized workloads. No ECC, 350W, and a 3-slot card.
Is renting cheaper than buying?
Renting avoids the upfront hardware cost and lets you match spend to actual usage. A rented RTX 3090 at $0.150/hr only costs money while it's running, whereas buying ties up capital in hardware that keeps depreciating whether it's in use or not. Buying outright runs $1,499 (launch MSRP). Which is cheaper depends on how continuously you'd run it; short or bursty workloads usually favor renting.
How is the RTX 3090 price calculated?
All prices are normalized to a per-GPU hourly rate using each offer's authoritative GPU count, which the raw price is divided by; some offers report price as already per-GPU. Offers whose price can't be safely normalized, or whose rate is an extreme outlier against the rest of the market, are excluded.
RTX 3090 price by region
Related guides
Other models in the same generation, then the rest of the GPU index.
