Home / Hardware guides / GPU rendering benchmark ladder
Data checked:
BLENDER 4.5.0 / CYCLES
GPU rendering rankings: Blender 4.5.0 benchmarks
Compare rendering scores, sample counts and compute backends in one place.
161Rendering test configurations
These are medians of public submissions, not measurements by LessWork Lab. Rendering scores do not represent gaming FPS.
TOM’S HARDWARE · NATIVE FPS
Gaming benchmark rankings
Compare native-resolution gaming FPS for 48 GPUs tested by Tom’s Hardware on the same platform. Switch resolution and rasterization or ray tracing to compare the published suite results.
These are Tom’s Hardware measurements, organized by LessWork Lab. No upscaling or frame generation is included. The table labels the presets Ultra; individual games use a mix of High and Ultra settings.
The default ranking below is 1440p Ultra rasterization. Open the other configurations to read every published result without JavaScript.
1440p Ultra · Rasterization
Showing 48 of 48 GPUs · 48 with reported FPS
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA167.3native FPS
#2GeForce RTX 4090NVIDIA143.4native FPS
#3GeForce RTX 5080NVIDIA128.3native FPS
#4Radeon RX 7900 XTXAMD122.3native FPS
#5GeForce RTX 4080 SuperNVIDIA118.6native FPS
#6GeForce RTX 4080NVIDIA117.5native FPS
#7GeForce RTX 5070 TiNVIDIA116.8native FPS
#8Radeon RX 9070 XTAMD116.5native FPS
#9Radeon RX 7900 XTAMD108.0native FPS
#10GeForce RTX 4070 Ti SuperNVIDIA104.0native FPS
#10Radeon RX 9070AMD104.0native FPS
#12GeForce RTX 3090 TiNVIDIA99.9native FPS
#13GeForce RTX 4070 TiNVIDIA97.9native FPS
#14GeForce RTX 5070NVIDIA96.4native FPS
#15GeForce RTX 3090NVIDIA91.5native FPS
#16GeForce RTX 4070 SuperNVIDIA91.2native FPS
#17Radeon RX 6950 XTAMD89.5native FPS
#18GeForce RTX 3080 TiNVIDIA89.1native FPS
#19Radeon RX 9070 GREAMD86.6native FPS
#20Radeon RX 7800 XTAMD84.7native FPS
#21Radeon RX 6900 XTAMD83.9native FPS
#22GeForce RTX 3080NVIDIA82.0native FPS
#23Radeon RX 6800 XTAMD79.6native FPS
#24GeForce RTX 4070NVIDIA77.8native FPS
#25GeForce RTX 5060 Ti 16GBNVIDIA73.4native FPS
#26Radeon RX 7700 XTAMD72.7native FPS
#27GeForce RTX 5060 Ti 8GBNVIDIA68.6native FPS
#28Radeon RX 9060 XT 16GBAMD67.3native FPS
#29GeForce RTX 3070 TiNVIDIA66.9native FPS
#30Radeon RX 9060 XT 8GBAMD62.5native FPS
#31GeForce RTX 4060 Ti 16GBNVIDIA60.5native FPS
#32GeForce RTX 5060NVIDIA59.8native FPS
#33GeForce RTX 4060 Ti 8GBNVIDIA58.9native FPS
#34GeForce RTX 3070NVIDIA58.2native FPS
#35Radeon RX 6750 XTAMD57.6native FPS
#36Radeon RX 6700 XTAMD54.3native FPS
#37GeForce RTX 3060 TiNVIDIA51.0native FPS
#38Intel Arc B580Intel50.7native FPS
#39Radeon RX 7600 XTAMD50.2native FPS
#40GeForce RTX 4060NVIDIA47.6native FPS
#41Radeon RX 7600AMD45.4native FPS
#42GeForce RTX 5050NVIDIA45.2native FPS
#43Intel Arc B570Intel44.3native FPS
#44GeForce RTX 3060 12GBNVIDIA41.9native FPS
#45Radeon RX 6600 XTAMD40.7native FPS
#46Radeon RX 6650 XTAMD38.0native FPS
#47GeForce RTX 3050NVIDIA29.8native FPS
#48Radeon RX 6600AMD24.8native FPS
No matching GPU. Change the search or brand filter.
Method, test platform and limitations
Rankings are calculated separately for the selected mode and resolution, using the published FPS in descending order. Equal published values share a competition rank; the following rank skips the tied places. Brand and search filters retain the full-list rank. Missing results are unranked rather than treated as zero. FPS retains the source’s one-decimal precision.
Rasterization uses the geometric mean of 11 games, with each game equally weighted. Ray tracing shows the source’s summary FPS for 8 tests; its captured RT section does not separately state the aggregation formula. A suite result does not predict FPS in an individual game.
Source test platform: CPU AMD Ryzen 7 9800X3D; memory G.Skill Trident Z5 32GB (2x16GB) DDR5-6000 CL30. Some tested cards have factory overclocks, so board models can differ from reference clocks.
This is a selected set of tested GPUs, not every GPU on the market. Results depend on the games, settings and CPU platform; RT tests include workloads with widely different ray/path tracing demands. The captured summary does not disclose individual driver versions, game patch versions, operating system or exact test dates for every GPU. Rounded ties do not prove identical performance or statistical significance. Price, power use and 1% lows are not ranked here.
The source update date describes the source page. Our check date records when we verified these published values; it does not make the underlying tests measurements from that later date.
Rasterization suite · 11 games
Black Myth Wukong · Alan Wake II · Apex Legends · Counter-Strike 2 · Fortnite · Arc Raiders · Stalker 2 · Marvel's Spider-Man 2 · Marvel Rivals · Cyberpunk 2077 · Clair Obscur: Expedition 33
Ray tracing suite · 8 games
Black Myth Wukong · Alan Wake II · DOOM: The Dark Ages · Grand Theft Auto V Enhanced · Marvel's Spider-Man 2 · Indiana Jones and the Great Circle · Assassin's Creed Shadows · Cyberpunk 2077
Other resolutions and ray tracing1080p Ultra · Rasterization
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA203.8native FPS
#2GeForce RTX 4090NVIDIA183.6native FPS
#3GeForce RTX 5080NVIDIA166.9native FPS
#4Radeon RX 7900 XTXAMD161.5native FPS
#5GeForce RTX 4080 SuperNVIDIA158.9native FPS
#6GeForce RTX 4080NVIDIA157.3native FPS
#7Radeon RX 9070 XTAMD156.6native FPS
#8GeForce RTX 5070 TiNVIDIA155.4native FPS
#9Radeon RX 7900 XTAMD145.4native FPS
#10GeForce RTX 4070 Ti SuperNVIDIA141.2native FPS
#11Radeon RX 9070AMD140.9native FPS
#12GeForce RTX 4070 TiNVIDIA135.1native FPS
#13GeForce RTX 5070NVIDIA132.6native FPS
#14GeForce RTX 3090 TiNVIDIA131.7native FPS
#15GeForce RTX 4070 SuperNVIDIA126.7native FPS
#16Radeon RX 6950 XTAMD123.3native FPS
#17GeForce RTX 3090NVIDIA122.9native FPS
#18Radeon RX 9070 GREAMD120.6native FPS
#19GeForce RTX 3080 TiNVIDIA119.6native FPS
#20Radeon RX 7800 XTAMD118.4native FPS
#21Radeon RX 6900 XTAMD117.1native FPS
#22Radeon RX 6800 XTAMD111.8native FPS
#23GeForce RTX 3080NVIDIA111.6native FPS
#24GeForce RTX 4070NVIDIA111.5native FPS
#25GeForce RTX 5060 Ti 16GBNVIDIA105.2native FPS
#26Radeon RX 7700 XTAMD102.9native FPS
#27Radeon RX 7600 XTAMD102.1native FPS
#28GeForce RTX 5060 Ti 8GBNVIDIA100.4native FPS
#29Radeon RX 9060 XT 16GBAMD98.3native FPS
#30GeForce RTX 3070 TiNVIDIA94.6native FPS
#31Radeon RX 9060 XT 8GBAMD93.2native FPS
#32GeForce RTX 4060 Ti 16GBNVIDIA89.3native FPS
#33GeForce RTX 5060NVIDIA88.5native FPS
#34GeForce RTX 4060 Ti 8GBNVIDIA88.0native FPS
#35GeForce RTX 3070NVIDIA87.2native FPS
#36Radeon RX 6750 XTAMD83.2native FPS
#37Radeon RX 6700 XTAMD79.2native FPS
#38GeForce RTX 3060 TiNVIDIA74.2native FPS
#39GeForce RTX 4060NVIDIA71.5native FPS
#39Intel Arc B580Intel71.5native FPS
#41Radeon RX 7600AMD69.9native FPS
#42GeForce RTX 5050NVIDIA69.3native FPS
#43Radeon RX 6650 XTAMD64.3native FPS
#44Intel Arc B570Intel63.5native FPS
#45Radeon RX 6600 XTAMD62.7native FPS
#46GeForce RTX 3060 12GBNVIDIA61.5native FPS
#47Radeon RX 6600AMD51.9native FPS
#48GeForce RTX 3050NVIDIA44.6native FPS
4K Ultra · Rasterization
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA110.8native FPS
#2GeForce RTX 4090NVIDIA89.1native FPS
#3GeForce RTX 5080NVIDIA77.4native FPS
#4Radeon RX 7900 XTXAMD70.6native FPS
#5GeForce RTX 4080 SuperNVIDIA69.4native FPS
#6GeForce RTX 5070 TiNVIDIA68.6native FPS
#7GeForce RTX 4080NVIDIA67.5native FPS
#8Radeon RX 9070 XTAMD65.8native FPS
#9Radeon RX 7900 XTAMD59.8native FPS
#10GeForce RTX 3090 TiNVIDIA59.3native FPS
#11GeForce RTX 4070 Ti SuperNVIDIA58.5native FPS
#12Radeon RX 9070AMD57.7native FPS
#13GeForce RTX 5070NVIDIA54.3native FPS
#14GeForce RTX 4070 TiNVIDIA53.9native FPS
#15GeForce RTX 3090NVIDIA53.1native FPS
#16GeForce RTX 3080 TiNVIDIA51.0native FPS
#17GeForce RTX 4070 SuperNVIDIA49.2native FPS
#18Radeon RX 6950 XTAMD48.3native FPS
#19Radeon RX 9070 GREAMD46.3native FPS
#20Radeon RX 7800 XTAMD45.1native FPS
#21Radeon RX 6900 XTAMD44.9native FPS
#22GeForce RTX 3080NVIDIA43.9native FPS
#23Radeon RX 6800 XTAMD42.2native FPS
#24GeForce RTX 4070NVIDIA41.3native FPS
#25GeForce RTX 5060 Ti 16GBNVIDIA40.2native FPS
#26Radeon RX 7700 XTAMD38.0native FPS
#27Radeon RX 9060 XT 16GBAMD35.2native FPS
#28GeForce RTX 4060 Ti 16GBNVIDIA31.3native FPS
#29Radeon RX 9060 XT 8GBAMD29.7native FPS
#30Radeon RX 6750 XTAMD29.6native FPS
#31GeForce RTX 5060 Ti 8GBNVIDIA28.1native FPS
#32Radeon RX 6700 XTAMD28.0native FPS
#33Intel Arc B580Intel27.6native FPS
#34GeForce RTX 3070NVIDIA26.2native FPS
#35GeForce RTX 3070 TiNVIDIA26.0native FPS
#36Radeon RX 7600 XTAMD25.5native FPS
#37GeForce RTX 4060 Ti 8GBNVIDIA23.9native FPS
#38GeForce RTX 3060 12GBNVIDIA22.2native FPS
#39GeForce RTX 5060NVIDIA21.8native FPS
#40Intel Arc B570Intel19.6native FPS
#41GeForce RTX 3060 TiNVIDIA19.4native FPS
#42Radeon RX 6650 XTAMD19.0native FPS
#43Radeon RX 7600AMD18.4native FPS
#44GeForce RTX 4060NVIDIA17.4native FPS
#45Radeon RX 6600 XTAMD17.3native FPS
#46GeForce RTX 5050NVIDIA17.0native FPS
#47Radeon RX 6600AMD14.5native FPS
#48GeForce RTX 3050NVIDIA12.6native FPS
1080p Ultra · Ray tracing
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA125.7native FPS
#2GeForce RTX 4090NVIDIA112.3native FPS
#3GeForce RTX 5080NVIDIA98.3native FPS
#4GeForce RTX 4080 SuperNVIDIA94.3native FPS
#5GeForce RTX 4080NVIDIA92.8native FPS
#6GeForce RTX 5070 TiNVIDIA90.4native FPS
#7GeForce RTX 4070 Ti SuperNVIDIA87.9native FPS
#8GeForce RTX 4070 TiNVIDIA78.5native FPS
#9Radeon RX 9070 XTAMD77.8native FPS
#10GeForce RTX 5070NVIDIA74.1native FPS
#11GeForce RTX 4070 SuperNVIDIA72.5native FPS
#12GeForce RTX 3090 TiNVIDIA72.3native FPS
#13Radeon RX 7900 XTXAMD70.4native FPS
#14Radeon RX 9070AMD67.6native FPS
#15GeForce RTX 3090NVIDIA66.5native FPS
#16GeForce RTX 3080 TiNVIDIA64.6native FPS
#17GeForce RTX 4070NVIDIA63.8native FPS
#18Radeon RX 7900 XTAMD61.7native FPS
#19Radeon RX 9070 GREAMD58.4native FPS
#20GeForce RTX 5060 Ti 16GBNVIDIA56.2native FPS
#21GeForce RTX 3080NVIDIA51.3native FPS
#22Radeon RX 7800 XTAMD48.4native FPS
#23GeForce RTX 4060 Ti 16GBNVIDIA48.3native FPS
#24Radeon RX 6950 XTAMD46.9native FPS
#25GeForce RTX 5060 Ti 8GBNVIDIA46.0native FPS
#26Radeon RX 9060 XT 16GBAMD45.3native FPS
#27Radeon RX 6900 XTAMD45.2native FPS
#28GeForce RTX 3070 TiNVIDIA42.1native FPS
#29Radeon RX 6800 XTAMD42.0native FPS
#30Radeon RX 7700 XTAMD41.6native FPS
#31GeForce RTX 5060NVIDIA40.9native FPS
#32GeForce RTX 4060 Ti 8GBNVIDIA40.8native FPS
#33GeForce RTX 3070NVIDIA39.7native FPS
#34Intel Arc B580Intel34.9native FPS
#35GeForce RTX 4060NVIDIA33.6native FPS
#36GeForce RTX 3060 TiNVIDIA33.2native FPS
#37Radeon RX 9060 XT 8GBAMD33.0native FPS
#38GeForce RTX 5050NVIDIA31.2native FPS
#39GeForce RTX 3060 12GBNVIDIA30.4native FPS
#40Radeon RX 6750 XTAMD28.5native FPS
#41Radeon RX 6700 XTAMD27.1native FPS
#42Radeon RX 7600 XTAMD27.0native FPS
#43Intel Arc B570Intel19.8native FPS
#44GeForce RTX 3050NVIDIA19.6native FPS
#45Radeon RX 6600 XTAMD17.2native FPS
#46Radeon RX 7600AMD16.0native FPS
#47Radeon RX 6650 XTAMD15.5native FPS
#48Radeon RX 6600AMD12.8native FPS
1440p Ultra · Ray tracing
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA100.8native FPS
#2GeForce RTX 4090NVIDIA87.7native FPS
#3GeForce RTX 5080NVIDIA74.0native FPS
#4GeForce RTX 4080 SuperNVIDIA69.5native FPS
#5GeForce RTX 4080NVIDIA68.4native FPS
#6GeForce RTX 5070 TiNVIDIA66.2native FPS
#7GeForce RTX 4070 Ti SuperNVIDIA60.0native FPS
#8Radeon RX 9070 XTAMD55.5native FPS
#9GeForce RTX 4070 TiNVIDIA55.3native FPS
#10GeForce RTX 3090 TiNVIDIA52.2native FPS
#11GeForce RTX 5070NVIDIA51.5native FPS
#12Radeon RX 7900 XTXAMD50.5native FPS
#13Radeon RX 9070AMD47.6native FPS
#14GeForce RTX 3090NVIDIA47.4native FPS
#15GeForce RTX 3080 TiNVIDIA46.2native FPS
#16GeForce RTX 4070 SuperNVIDIA43.6native FPS
#17Radeon RX 7900 XTAMD43.5native FPS
#18GeForce RTX 5060 Ti 16GBNVIDIA38.9native FPS
#19GeForce RTX 4070NVIDIA38.6native FPS
#20Radeon RX 9070 GREAMD37.3native FPS
#21GeForce RTX 3080NVIDIA34.7native FPS
#22Radeon RX 7800 XTAMD33.5native FPS
#23Radeon RX 6950 XTAMD32.8native FPS
#24GeForce RTX 4060 Ti 16GBNVIDIA32.7native FPS
#25Radeon RX 6900 XTAMD32.1native FPS
#26Radeon RX 9060 XT 16GBAMD31.0native FPS
#27Radeon RX 6800 XTAMD29.6native FPS
#28GeForce RTX 3070 TiNVIDIA27.7native FPS
#29GeForce RTX 5060 Ti 8GBNVIDIA26.6native FPS
#30GeForce RTX 3070NVIDIA26.2native FPS
#31GeForce RTX 4060 Ti 8GBNVIDIA25.7native FPS
#32Radeon RX 7700 XTAMD25.3native FPS
#33GeForce RTX 5060NVIDIA24.6native FPS
#34Intel Arc B580Intel23.1native FPS
#35GeForce RTX 3060 TiNVIDIA22.3native FPS
#36GeForce RTX 4060NVIDIA21.3native FPS
#37GeForce RTX 3060 12GBNVIDIA20.7native FPS
#38Radeon RX 9060 XT 8GBAMD20.2native FPS
#39Radeon RX 6750 XTAMD19.1native FPS
#40GeForce RTX 5050NVIDIA18.7native FPS
#41Radeon RX 7600 XTAMD18.1native FPS
#42Radeon RX 6700 XTAMD18.0native FPS
#43Intel Arc B570Intel13.1native FPS
#44GeForce RTX 3050NVIDIA12.6native FPS
#45Radeon RX 7600AMD9.3native FPS
#46Radeon RX 6650 XTAMD9.2native FPS
#47Radeon RX 6600 XTAMD9.1native FPS
#48Radeon RX 6600AMD7.7native FPS
4K Ultra · Ray tracing
RankGPU modelSuite FPS
#1GeForce RTX 5090NVIDIA64.1native FPS
#2GeForce RTX 4090NVIDIA52.2native FPS
#3GeForce RTX 5080NVIDIA42.1native FPS
#4GeForce RTX 4080 SuperNVIDIA38.5native FPS
#5GeForce RTX 4080NVIDIA37.4native FPS
#6GeForce RTX 5070 TiNVIDIA36.7native FPS
#7GeForce RTX 4070 Ti SuperNVIDIA32.2native FPS
#8Radeon RX 9070 XTAMD30.1native FPS
#9GeForce RTX 3090 TiNVIDIA29.4native FPS
#10Radeon RX 7900 XTXAMD27.3native FPS
#11GeForce RTX 3090NVIDIA26.3native FPS
#12Radeon RX 9070AMD25.8native FPS
#13GeForce RTX 4070 TiNVIDIA23.9native FPS
#14Radeon RX 7900 XTAMD23.1native FPS
#15GeForce RTX 5070NVIDIA22.7native FPS
#16GeForce RTX 4070 SuperNVIDIA22.1native FPS
#17GeForce RTX 5060 Ti 16GBNVIDIA20.6native FPS
#18GeForce RTX 4070NVIDIA19.3native FPS
#19GeForce RTX 3080 TiNVIDIA18.0native FPS
#20GeForce RTX 4060 Ti 16GBNVIDIA16.8native FPS
#21Radeon RX 6950 XTAMD15.6native FPS
#22Radeon RX 9070 GREAMD15.5native FPS
#23Radeon RX 6900 XTAMD15.1native FPS
#23Radeon RX 7800 XTAMD15.1native FPS
#25Radeon RX 9060 XT 16GBAMD14.9native FPS
#26Radeon RX 6800 XTAMD14.1native FPS
#27GeForce RTX 3080NVIDIA13.8native FPS
#28Radeon RX 7700 XTAMD9.8native FPS
#29GeForce RTX 3060 12GBNVIDIA9.5native FPS
#30GeForce RTX 3070 TiNVIDIA8.8native FPS
#31Radeon RX 7600 XTAMD8.5native FPS
#32GeForce RTX 3070NVIDIA8.3native FPS
#33GeForce RTX 4060 Ti 8GBNVIDIA8.1native FPS
#33Radeon RX 6750 XTAMD8.1native FPS
#35Radeon RX 6700 XTAMD7.9native FPS
#36GeForce RTX 3060 TiNVIDIA7.4native FPS
#37Radeon RX 9060 XT 8GBAMD7.3native FPS
#38GeForce RTX 5060 Ti 8GBNVIDIA7.2native FPS
#39GeForce RTX 4060NVIDIA6.8native FPS
#39GeForce RTX 5060NVIDIA6.8native FPS
#41GeForce RTX 5050NVIDIA6.2native FPS
#42GeForce RTX 3050NVIDIA4.5native FPS
#43Radeon RX 7600AMD4.1native FPS
#44Intel Arc B580Intel3.8native FPS
#45Radeon RX 6600 XTAMD3.4native FPS
#45Radeon RX 6650 XTAMD3.4native FPS
#47Intel Arc B570Intel3.3native FPS
#48Radeon RX 6600AMD3.1native FPS
One test, comparable relative performance.
Blender 4.5.0 · Cycles rendering · Median of public submissions · Higher is faster. These are rendering configuration rankings, not gaming FPS or gaming GPU rankings.
161 records included; search and filter available.
NVIDIA / GeForceRendering scoreAMD / Radeon
GeForce RTX 5090 D
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 12 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1553 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 21 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 526 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 4 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1431 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 238 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 219 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1415 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 253 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 156 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 463 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 914 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 139 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 629 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 208 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 262 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 579 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 223 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 655 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 14 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 332 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 372 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 120 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 117 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 676 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 312 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 439 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 44 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 59 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 90 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 136 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 232 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 35 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 56 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 98 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 42 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 10 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 40 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 41 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 104 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 630 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 66 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 24 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 109 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 36 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 4 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 213 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 168 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 9 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 23 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 85 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 4 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 145 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 31 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 7 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 75 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 8 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 24 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 39 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 51 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 33 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 89 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 63 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 112 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 39 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 62 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 21 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 9 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 22 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 25 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 81 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 19 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 78 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 10 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 31 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 146 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 8 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 62 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 17 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 9 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 33 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 22 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 10 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · HIP backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 73 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 36 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 17 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 6 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 10 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 5 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 3 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 22 public submissions.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 2 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · OPTIX backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
Blender 4.5.0 Cycles · CUDA backend. This score is the median of 1 public submissions. Fewer than 5 submissions: pay particular attention to the sample limitation.
These records retain the original device name and backend. Platforms, power and cooling are not standardized; they are not tests conducted by this site.
No matching test configurations. Change your search criteria.
Scores, rankings and data limitations
Source: Blender Foundation’s Blender Open Data (CC0). LessWork Lab organizes public submissions; we did not test this hardware. We use the official median for exact Blender 4.5.0, grouped by original device name and compute backend, across operating systems. The score sums estimated Cycles samples per minute across benchmark scenes; higher is faster for this rendering workload. Rankings use the unrounded numeric median; equal original medians share a rank. Original device/backend pairs remain separate. Counts are public submissions, not independent machines or people. Drivers, power, cooling and hardware settings are not standardized; each submission’s completion of all scenes has not been independently verified here.
GPU rankings: first check what the benchmark measures
Average gaming FPS, ray tracing tests, synthetic graphics tests and offline rendering throughput measure different things. Resolution, quality settings, driver version and CPU configuration also affect results. Raw scores from different tests do not share a unit. Even percentages require the same reference card and test conditions.
The rendering tab uses medians of public Blender 4.5.0 Cycles submissions and retains original device names, backends and submission counts. The results describe rendering in those configurations; they cannot be converted into gaming FPS or used as ray tracing game or AI inference rankings. NVIDIA OptiX/CUDA and AMD HIP are shown in the entries. Different backends are different test configurations and must not be silently combined. The data comes from public sources; this site does not claim to have tested all products itself.
Small samples are more sensitive to individual submissions. Overclocking, cooling, power limits and system settings also cause differences, so close scores do not establish a permanent winner. A rank applies only to the devices and configurations included in that tab, not to every graphics card on the market.
Rendering benchmark questions
Can Blender scores serve as gaming GPU rankings?
They cannot directly replace gaming tests. Blender measures Cycles rendering throughput; games depend on their engine, resolution, ray tracing, CPU and quality settings. Read the two rankings separately.
What does the update time at the top mean?
It is the actual verification time for this data batch. New submissions may appear at the source, but refreshing your browser does not change this page’s timestamp. It changes only after data is fetched again, its methodology is verified and the page is updated.
Rendering data sources
Check the original device name, compute backend and sample count together. The timestamp above identifies this verified snapshot; it does not change when you refresh the page.
This page retains 161 previously selected GPU configurations and their exact device/backend pairs, rather than all 493 configurations in the checked official snapshot. 63 selected configurations have fewer than five public submissions and remain ranked with a caution label. The CPU page instead separates configurations with fewer than five submissions from its main rankings. This difference is an editorial display policy, not an official Blender confidence rule.
Source: Blender Foundation’s Blender Open Data (CC0). LessWork Lab organizes public submissions; we did not test this hardware. We use the official median for exact Blender 4.5.0, grouped by original device name and compute backend, across operating systems. The score sums estimated Cycles samples per minute across benchmark scenes; higher is faster for this rendering workload. Rankings use the unrounded numeric median; equal original medians share a rank. Original device/backend pairs remain separate. Counts are public submissions, not independent machines or people. Drivers, power, cooling and hardware settings are not standardized; each submission’s completion of all scenes has not been independently verified here.
CPU and GPU sources are scheduled for checks every three days. The displayed data date changes only after source verification, independent review and publication succeed; failed checks retain the last verified version.
GPU rendering ranking questions
Can these GPU rankings predict game FPS or AI speed?
No. Blender Cycles measures rendering throughput. Game FPS, AI inference, power efficiency and value require measurements for those uses.
Why are low-submission GPUs still ranked?
We retain the previously selected GPU configurations and flag those with fewer than five submissions. CPU configurations below that threshold are shown separately. Neither policy is a Blender confidence classification; submission counts are not counts of independent machines or people.
Why can the same GPU have more than one row?
Rows preserve exact device names and compute backends such as OptiX and CUDA. Different backends are separate score groups; they are not merged into a single hardware score.