48score
#45 of 131
Overall rank
NVIDIAMid-range
TITAN RTX
As an Amazon Associate we may earn from qualifying purchases.
VRAM
24 GB
CUDA
4,608
FP32
16.31 TF
Bandwidth
672 GB/s
TDP
280W
Boost
1770 MHz
Strengths at a Glance
How it stacks up to the flagship
Each metric is shown as a percentage of the GeForce RTX 5090 D, the strongest card we track.
FP32 compute16.31 TFLOPS16%
vs RTX 5090 D: 104.8 TFLOPS
Memory bandwidth672 GB/s38%
vs RTX 5090 D: 1790 GB/s
VRAM capacity24 GB75%
vs RTX 5090 D: 32 GB
Shading units4,60821%
vs RTX 5090 D: 21,760
Power efficiency19/10033%
vs RTX 5090 D: 58/100
Synthetic Benchmarks
Estimated benchmark results
Each result is shown as a share of the RTX 5090 D's score in the same test.
3DMark Time Spy17,280pts
vs RTX 5090 D: 36,000 pts
3DMark Port Royal (RT)8,820pts
vs RTX 5090 D: 18,000 pts
Blender (samples/min)2,496spm
vs RTX 5090 D: 5,200 spm
Geekbench Compute120,000pts
vs RTX 5090 D: 240,000 pts
Performance breakdown
Gaming48
Ray tracing49
AI / Compute50
Creator / 3D48
Power efficiency19
Real-World Gaming
FPS Across Resolutions
1080p · Ultra preset1440p · Ultra preset4K · Ultra preset
Cyberpunk 2077avg 57 fps
1080p
76
1440p
64
4K
32
Counter-Strike 2avg 157 fps
1080p
262
1440p
152
4K
56
Fortniteavg 121 fps
1080p
171
1440p
121
4K
72
Battlefield 5avg 120 fps
1080p
155
1440p
112
4K
94
Far Cry 5avg 117 fps
1080p
142
1440p
130
4K
78
Valorantavg 273 fps
1080p
229
1440p
298
4K
291
Average FPS across all PC games
1080p
172fps
1440p
107fps
4K
77fps
Cost per frame
Launch MSRP ($2499) ÷ average FPS — lower is better.
1080p
$14.71
1440p
$22.79
4K
$35.99
Community Feedback
What Owners Say
Owners love the raw compute power for pro work and AI models. The common gripe is the insane price and how loud the single fan gets under load.
Pros
- Massive VRAM for giant projects
- Crushes 4K video and rendering
- Runs cool under sustained load
- Professional driver support for creators
Cons
- Noisy under heavy gaming load
- Extremely expensive for what it is
- No hardware ray tracing performance uplift
Supported technologies
Ray TracingDLSSNVENCAV1 Encode
Full specifications
Graphics processor
Core configuration
- CUDA Cores
- 4,608
- RT Cores
- 72
- Tensor Cores
- 576
- TMUs
- 288
- ROPs
- 96
- L2 cache
- 6 MB
Memory
- Size
- 24 GB
- Type
- GDDR6
- Bus width
- 384-bit
- Bandwidth
- 672 GB/s
- Memory clock
- 1750 MHz
Clocks & throughput
- Base clock
- 1350 MHz
- Boost clock
- 1770 MHz
- FP32 (float)
- 16.31 TFLOPS
- FP16 (half)
- 16.31 TFLOPS
- Pixel rate
- 170 GPixel/s
- Texture rate
- 509.8 GTexel/s
Board & power
- TDP
- 280W
- Suggested PSU
- 550W
- Power connectors
- 2x 8-pin
- Bus interface
- PCIe 3.0 x16
- Length
- 267 mm
- Slot width
- 2-slot
Display & outputs
- Max resolution
- 7680×4320
- Outputs
- 1x HDMI, 3x DisplayPort, 1x USB Type-C
API and SDK support
- DirectX
- 12 Ultimate (12_1)
- Shader Model
- 6.5
- OpenGL
- 4.6
- OpenCL
- 2.0
- Vulkan
- 1.2.131
- CUDA
- 7.5
Verdict
Our verdict on the TITAN RTX
The NVIDIA TITAN RTX is a workstation-focused card built for pro creative tasks, with its massive memory capacity being the main draw.
Get it if you’re a pro doing heavy 3D rendering or AI training and need all the VRAM you can get. Skip it if you’re gaming—cheaper cards match or beat it in games for way less.
Buy it if…
- You need maximum VRAM for massive AI or 3D datasets.
- You run heavy multi-GPU render farms without memory limits.
- You want a workstation card that also crushes 4K gaming.
Leaderboard
Its place in the overall top
3.5
106 votes
Rate this GPU
Add your verdict
Keep exploring