G3D Mark is a standard benchmark that measures graphics performance in real-world gaming scenarios. It simplifies comparing cards from different brands, where higher scores directly correlate with better fps and smoother gaming experiences.

Head-to-Head Verdict, Benchmarks, Value & Long-Term Outlook

This comparison brings together gaming FPS, raw graphics performance, VRAM, feature set, power efficiency, pricing context, and long-term value so you can see which GPU actually makes more sense.

GeForce RTX 5090

2025

Why buy it

✅+19.2% higher PassMark G3D performance.
✅Costs $29,001 less on MSRP ($1,999 MSRP vs $31,000 MSRP).
✅Delivers 1748.8% more G3D Mark for each dollar spent, at 19.4 vs 1.1 G3D/$ ($1,999 MSRP vs $31,000 MSRP).
✅Access to a newer frame-generation stack with DLSS 4 Multi Frame Generation (2025).
✅More future proof: Blackwell 2.0 (2025−2026) on 5nm with a newer platform for upcoming games.

Trade-offs

❌Less VRAM, with 32 GB vs 48 GB for high-resolution textures and newer games.
❌91.7% higher power demand at 575W vs 300W.
❌13.9% longer card at 304mm vs 267mm.

L40

2022

Why buy it

✅50% more VRAM for high-resolution textures and newer games (48 GB vs 32 GB).
✅Draws 300W instead of 575W, a 275W reduction.
✅Measures 267mm instead of 304mm, a 37mm shorter card that is more SFF-friendly.

Trade-offs

❌Lower PassMark G3D performance (32,601 vs 38,867).
❌No equivalent frame-generation stack like DLSS 4 Multi Frame Generation (2025).
❌Weaker long-term outlook: GeForce RTX 5090 is the safer future-proof pick thanks to newer hardware and better gaming feature support.
❌
1450.8% HIGHER MSRP
$31,000 MSRPvs$1,999 MSRP
❌Lower G3D Mark per dollar, at 1.1 vs 19.4 G3D/$ ($31,000 MSRP vs $1,999 MSRP).

GeForce RTX 5090

2025

L40

2022

Why buy it

✅+19.2% higher PassMark G3D performance.
✅Costs $29,001 less on MSRP ($1,999 MSRP vs $31,000 MSRP).
✅Delivers 1748.8% more G3D Mark for each dollar spent, at 19.4 vs 1.1 G3D/$ ($1,999 MSRP vs $31,000 MSRP).
✅Access to a newer frame-generation stack with DLSS 4 Multi Frame Generation (2025).
✅More future proof: Blackwell 2.0 (2025−2026) on 5nm with a newer platform for upcoming games.

Why buy it

✅50% more VRAM for high-resolution textures and newer games (48 GB vs 32 GB).
✅Draws 300W instead of 575W, a 275W reduction.
✅Measures 267mm instead of 304mm, a 37mm shorter card that is more SFF-friendly.

Trade-offs

❌Less VRAM, with 32 GB vs 48 GB for high-resolution textures and newer games.
❌91.7% higher power demand at 575W vs 300W.
❌13.9% longer card at 304mm vs 267mm.

Trade-offs

❌Lower PassMark G3D performance (32,601 vs 38,867).
❌No equivalent frame-generation stack like DLSS 4 Multi Frame Generation (2025).
❌Weaker long-term outlook: GeForce RTX 5090 is the safer future-proof pick thanks to newer hardware and better gaming feature support.
❌
1450.8% HIGHER MSRP
$31,000 MSRPvs$1,999 MSRP
❌Lower G3D Mark per dollar, at 1.1 vs 19.4 G3D/$ ($31,000 MSRP vs $1,999 MSRP).

Quick Answers

So, is GeForce RTX 5090 better than L40?

Yes, but this is not really about a huge raw performance gap. The broader synthetic picture is also very close at 38,867 vs 32,601 in G3D Mark. The bigger reason to prefer GeForce RTX 5090 is the overall package: you are getting a newer generation, DLSS 4 + Multi Frame Gen.

Which one is more future-proof for 2026 and beyond?

GeForce RTX 5090 is the more future-proof choice for 2026 and beyond. You are getting a newer 2025 generation instead of 2022, 19.2% more raw performance headroom, and better upscaling support with DLSS 4 Super Resolution (2025) instead of no meaningful modern upscaling stack and better frame-generation support with DLSS 4 Multi Frame Generation (2025) instead of no meaningful modern upscaling stack. That broader feature stack should age better as more games lean on modern upscaling and frame-generation support.

Which one is the smarter buy today, not just the cheaper card?

GeForce RTX 5090 is the smarter buy today, but it is not as lopsided as a simple winner label makes it sound. GeForce RTX 5090 is about $29,001 cheaper on MSRP at $1,999 MSRP versus $31,000 MSRP, and you are getting 19.2% higher G3D Mark. It also leads G3D-per-dollar by 1748.8%. That is why the better overall card still comes out as the smarter buy today, not just the faster one.

Is L40 still worth buying for gaming in 2026?

Yes. L40 is still a strong modern gaming GPU: it is excellent for 1080p and 1440p maxed-out gaming, and still very capable for 4K in many titles. It remains a good buy when you can get it meaningfully cheaper than the alternative around $31,000 MSRP, even if GeForce RTX 5090 is still the cleaner recommendation on overall value today.

Games Benchmarks

Real-world benchmarks and performance projections based on comprehensive hardware analysis and comparative metrics. Values represent expected performance on High/Ultra settings at 1080p, 1440p, and 4K. Modeled using a Ryzen 7 9800X3D reference profile to minimize specific CPU bottlenecks.

Note: Performance behavior can vary per game. Specific architectures may perform better or worse depending on game engine optimizations and API implementation.

GeForce RTX 5090 Benchmark Check

L40 Benchmark Check

Path of Exile 2

Preset	GeForce RTX 5090	L40
1080p
low	276 FPS	281 FPS
medium	257 FPS	252 FPS
high	219 FPS	207 FPS
ultra	192 FPS	131 FPS
1440p
low	260 FPS	251 FPS
medium	217 FPS	211 FPS
high	172 FPS	152 FPS
ultra	155 FPS	95 FPS
4K
low	179 FPS	116 FPS
medium	149 FPS	97 FPS
high	112 FPS	66 FPS
ultra	100 FPS	57 FPS

Open GeForce RTX 5090 Open L40

Path of Exile 2

Path of Exile 2

Counter-Strike 2

Preset	GeForce RTX 5090	L40
1080p
low	880 FPS	525 FPS
medium	713 FPS	407 FPS
high	561 FPS	310 FPS
ultra	474 FPS	268 FPS
1440p
low	697 FPS	440 FPS
medium	591 FPS	351 FPS
high	483 FPS	273 FPS
ultra	383 FPS	215 FPS
4K
low	383 FPS	233 FPS
medium	328 FPS	183 FPS
high	305 FPS	162 FPS
ultra	263 FPS	134 FPS

Open GeForce RTX 5090 Open L40

Counter-Strike 2

Counter-Strike 2

League of Legends

Preset	GeForce RTX 5090	L40
1080p
low	944 FPS	999 FPS
medium	772 FPS	999 FPS
high	698 FPS	978 FPS
ultra	604 FPS	734 FPS
1440p
low	731 FPS	970 FPS
medium	598 FPS	880 FPS
high	529 FPS	734 FPS
ultra	452 FPS	550 FPS
4K
low	506 FPS	569 FPS
medium	419 FPS	493 FPS
high	377 FPS	437 FPS
ultra	316 FPS	367 FPS

Open GeForce RTX 5090 Open L40

League of Legends

League of Legends

Valorant

Preset	GeForce RTX 5090	L40
1080p
low	999 FPS	492 FPS
medium	999 FPS	430 FPS
high	907 FPS	383 FPS
ultra	811 FPS	293 FPS
1440p
low	885 FPS	346 FPS
medium	778 FPS	311 FPS
high	685 FPS	273 FPS
ultra	592 FPS	208 FPS
4K
low	639 FPS	229 FPS
medium	568 FPS	197 FPS
high	507 FPS	156 FPS
ultra	437 FPS	119 FPS

Open GeForce RTX 5090 Open L40

Valorant

Valorant

Check FPS for your specific build

Technical Specifications

Side-by-side comparison of GeForce RTX 5090 and L40

GeForce RTX 5090

The GeForce RTX 5090 is manufactured by NVIDIA. It was released in January 30 2025. It features the Blackwell 2.0 architecture. The core clock ranges from 2017 MHz to 2407 MHz. It has 21760 shading units. The thermal design power (TDP) is 575W. Manufactured using 5 nm process technology. It features 170 dedicated ray tracing cores for enhanced lighting effects. G3D Mark benchmark score: 38,867 points. Launch price was $1,999.

L40

The L40 is manufactured by NVIDIA. It was released in October 13 2022. It features the Ada Lovelace architecture. The core clock ranges from 735 MHz to 2490 MHz. It has 18176 shading units. The thermal design power (TDP) is 300W. Manufactured using 4 nm process technology. It features 142 dedicated ray tracing cores for enhanced lighting effects. G3D Mark benchmark score: 32,601 points.

⚡

Graphics Performance

In G3D Mark, the GeForce RTX 5090 scores 38,867 versus the L40's 32,601 — the GeForce RTX 5090 leads by 19.2%. The GeForce RTX 5090 is built on Blackwell 2.0 while the L40 uses Ada Lovelace, both on 5 nm vs 4 nm. Shader units: 21,760 (GeForce RTX 5090) vs 18,176 (L40). Raw compute: 104.8 TFLOPS (GeForce RTX 5090) vs 90.52 TFLOPS (L40). Boost clocks: 2407 MHz vs 2490 MHz. Ray tracing: 170 RT cores (GeForce RTX 5090) vs 142 (L40) with 680 Tensor cores vs 568.

Feature	GeForce RTX 5090	L40
G3D Mark Score	38,867+19%	32,601
Architecture	Blackwell 2.0	Ada Lovelace
Process Node	5 nm	4 nm
Shading Units	21760+20%	18176
Compute (TFLOPS)	104.8 TFLOPS+16%	90.52 TFLOPS
Boost Clock	2407 MHz	2490 MHz+3%
ROPs	176	192+9%
TMUs	680+20%	568
L1 Cache	21.3 MB+20%	17.8 MB
L2 Cache	96 MB	96 MB
Ray Tracing Cores	170+20%	142
Tensor Cores	680+20%	568

✨

Advanced Features (DLSS/FSR)

A critical advantage for the GeForce RTX 5090 is support for DLSS 4 Multi Frame Generation. This allows it to generate entire frames using AI/Algorithms, essentially doubling the frame rate in CPU-bound scenarios or heavy ray-tracing titles. The L40 lacks specific hardware/driver support for this native frame generation tier.The GeForce RTX 5090 gives access to NVIDIA DLSS (Deep Learning Super Sampling), widely regarding as the superior upscaling method for image quality. The L40 relies on FSR (FidelityFX Super Resolution), which is capable but generally slightly noisier than DLSS in motion.

Feature	GeForce RTX 5090	L40
Upscaling Tech	DLSS 4 Super Resolution	Upscaling support
Frame Generation	DLSS 4 Multi Frame Generation	Not Supported
Ray Reconstruction	Yes (DLSS 4)	No
Low Latency	NVIDIA Reflex	Standard

💾

Video Memory (VRAM)

The GeForce RTX 5090 comes with 32 GB of VRAM, while the L40 has 48 GB. The L40 offers 50% more capacity, crucial for higher resolutions and texture-heavy games. Memory bandwidth: 1792 GB/s (GeForce RTX 5090) vs 960 GB/s (L40) — a 86.7% advantage for the GeForce RTX 5090. Bus width: 512-bit vs 384-bit.

Feature	GeForce RTX 5090	L40
VRAM Capacity	32 GB	48 GB+50%
Memory Type	GDDR7	GDDR6
Memory Bandwidth	1792 GB/s+87%	960 GB/s
Bus Width	512-bit+33%	384-bit
L2 Cache	96 MB	96 MB

🖥️

Display & API Support

DirectX support: 12.2 (GeForce RTX 5090) vs 12.2 (L40). Vulkan: 1.4 vs 1.3. OpenGL: 4.6 vs 4.6. Maximum simultaneous displays: 4 vs 4.

Feature	GeForce RTX 5090	L40
DirectX	12.2	12.2
Vulkan	1.4+8%	1.3
OpenGL	4.6	4.6
Max Displays	4	4

🎬

Media & Encoding

Hardware encoder: 9th Gen NVENC (2x) (GeForce RTX 5090) vs NVENC 8th Gen (L40). Decoder: 6th Gen NVDEC vs NVDEC 5th Gen. Supported codecs: MPEG-2,H.264,HEVC,VP9,AV1 (GeForce RTX 5090) vs AV1,HEVC,H.264,VP9 (L40).

Feature	GeForce RTX 5090	L40
Encoder	9th Gen NVENC (2x)	NVENC 8th Gen
Decoder	6th Gen NVDEC	NVDEC 5th Gen
Codecs	MPEG-2,H.264,HEVC,VP9,AV1	AV1,HEVC,H.264,VP9

🔌

Power & Dimensions

The GeForce RTX 5090 draws 575W versus the L40's 300W — a 62.9% difference. The L40 is more power-efficient. Recommended PSU: 1000W (GeForce RTX 5090) vs 750W (L40). Power connectors: 16-pin (12VHPWR) vs 16-pin. Card length: 304mm vs 267mm, occupying 2 vs 2 slots. Typical load temperature: 85°C vs 80°C.

Feature	GeForce RTX 5090	L40
TDP	575W	300W-48%
Recommended PSU	1000W	750W-25%
Power Connector	16-pin (12VHPWR)	16-pin
Length	304mm	267mm
Height	137mm	111mm
Slots	2	2
Temp (Load)	85°C	80°C-6%
Perf/Watt	67.6	108.7+61%

💰

Value Analysis

The GeForce RTX 5090 launched at $1999 MSRP, while the L40 launched at $31000. The GeForce RTX 5090 costs 93.6% less ($29001 savings) on MSRP. Performance per dollar on MSRP (G3D Mark / MSRP): 19.4 (GeForce RTX 5090) vs 1.1 (L40) — the GeForce RTX 5090 offers 1663.6% better value. The GeForce RTX 5090 is the newer GPU (2025 vs 2022).