GPUDECK · MEASURED PAIR · NO WINNER

GeForce RTX 4080 vs GeForce RTX 5090, one workload at a time

These two cards ran the same paid production workloads on the same days, so their goodput can be compared as a ratio inside each workload cell. The ratio is stated as GeForce RTX 4080 over GeForce RTX 5090 and changes with the workload — which is the finding. There is no single number, no winner, and no ranking.

107 published cells · 14 observation day(s) (2026-08-21 → 2026-09-04) · 107 meet the replication floor · page generated 2026-09-23 05:21 UTC

Every published cell

WORKLOADDAYGOODPUT GeForce RTX 4080 / GeForce RTX 5090ATTEMPTS A/BCARDS A/BCARD-DAYS A/B
chroma-v48-detail-svd_fp82026-08-210.39×42/1552/62/6
chroma-v48-detail-svd_fp82026-08-220.25×25/1872/62/6
chroma-v48-detail-svd_fp82026-08-230.48×55/1572/62/6
chroma-v48-detail-svd_fp82026-08-240.48×40/912/42/4
chroma-v48-detail-svd_fp82026-08-260.35×79/1952/62/6
chroma-v48-detail-svd_fp82026-08-270.36×131/1782/62/6
chroma-v48-detail-svd_fp82026-08-290.36×54/312/32/3
chroma-v48-detail-svd_fp82026-09-010.29×79/682/22/2
chroma-v48-detail-svd_fp82026-09-030.51×66/272/42/4
chroma1-hd_fp8_scaled2026-08-210.34×65/1042/62/6
chroma1-hd_fp8_scaled2026-08-220.28×38/862/62/6
chroma1-hd_fp8_scaled2026-08-230.53×34/542/62/6
chroma1-hd_fp8_scaled2026-08-240.41×63/702/42/4
chroma1-hd_fp8_scaled2026-08-260.47×107/1242/62/6
chroma1-hd_fp8_scaled2026-08-270.44×156/1592/62/6
chroma1-hd_fp8_scaled2026-08-290.28×104/272/52/5
chroma1-hd_fp8_scaled2026-09-020.46×8/122/42/4
chroma1-hd_fp8_scaled2026-09-030.68×31/152/32/3
dark_beast_krea2_fp82026-08-210.37×308/13122/62/6
dark_beast_krea2_fp82026-08-220.33×205/13312/62/6
dark_beast_krea2_fp82026-08-230.38×406/16932/62/6
dark_beast_krea2_fp82026-08-240.59×76/2942/62/6
dark_beast_krea2_fp82026-08-260.38×393/26522/62/6
dark_beast_krea2_fp82026-08-270.36×389/30422/62/6
dark_beast_krea2_fp82026-08-280.39×54/1362/62/6
dark_beast_krea2_fp82026-08-290.40×360/15802/62/6
dark_beast_krea2_fp82026-08-300.53×29/2112/52/5
dark_beast_krea2_fp82026-08-310.34×99/16002/62/6
dark_beast_krea2_fp82026-09-010.32×48/3382/62/6
dark_beast_krea2_fp82026-09-020.41×23/1372/62/6
dark_beast_krea2_fp82026-09-030.97×754/3872/52/5
dark_beast_krea2_fp82026-09-040.56×101/1362/52/5
dark_beast_krea2_identity_edit_v1_22026-08-210.46×120/3902/62/6
dark_beast_krea2_identity_edit_v1_22026-08-220.44×177/5182/62/6
dark_beast_krea2_identity_edit_v1_22026-08-230.43×131/4172/62/6
dark_beast_krea2_identity_edit_v1_22026-08-240.44×144/582/62/6
dark_beast_krea2_identity_edit_v1_22026-08-260.42×262/6262/62/6
dark_beast_krea2_identity_edit_v1_22026-08-270.44×330/7952/62/6
dark_beast_krea2_identity_edit_v1_22026-08-280.55×75/182/42/4
dark_beast_krea2_identity_edit_v1_22026-08-290.40×236/1412/62/6
dark_beast_krea2_identity_edit_v1_22026-08-300.52×9/162/32/3
dark_beast_krea2_identity_edit_v1_22026-08-310.35×119/332/32/3
dark_beast_krea2_identity_edit_v1_22026-09-010.34×66/392/32/3
dark_beast_krea2_identity_edit_v1_22026-09-020.49×18/292/62/6
dark_beast_krea2_identity_edit_v1_22026-09-030.76×670/1692/52/5
dark_beast_krea2_identity_edit_v1_22026-09-040.51×88/382/52/5
dark_beast_z_image_turbo_v9_bf162026-08-210.44×80/2902/62/6
dark_beast_z_image_turbo_v9_bf162026-08-220.34×113/4242/62/6
dark_beast_z_image_turbo_v9_bf162026-08-230.33×108/2952/62/6
dark_beast_z_image_turbo_v9_bf162026-08-240.30×95/2522/42/4
dark_beast_z_image_turbo_v9_bf162026-08-260.37×118/3082/62/6
dark_beast_z_image_turbo_v9_bf162026-08-270.29×92/3052/62/6
dark_beast_z_image_turbo_v9_bf162026-08-290.27×100/502/42/4
flux1-schnell-fp82026-08-270.25×18/582/52/5
krea2_identity_edit_sogni_v0_3_alpha2026-08-210.33×24/572/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-220.37×32/852/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-230.37×86/1532/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-240.51×18/342/52/5
krea2_identity_edit_sogni_v0_3_alpha2026-08-260.26×54/1612/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-270.43×38/1292/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-290.40×75/432/62/6
krea2_identity_edit_sogni_v0_3_alpha2026-08-310.54×29/202/22/2
krea2_identity_edit_sogni_v0_3_alpha2026-09-030.90×96/392/52/5
krea2_identity_edit_v1_22026-08-210.33×256/7622/62/6
krea2_identity_edit_v1_22026-08-220.34×244/8132/62/6
krea2_identity_edit_v1_22026-08-230.42×165/7182/62/6
krea2_identity_edit_v1_22026-08-240.56×41/1232/62/6
krea2_identity_edit_v1_22026-08-260.44×370/7332/62/6
krea2_identity_edit_v1_22026-08-270.38×368/6632/62/6
krea2_identity_edit_v1_22026-08-290.38×377/2032/62/6
krea2_identity_edit_v1_22026-08-310.37×195/212/22/2
krea2_identity_edit_v1_22026-09-020.79×38/112/42/4
krea2_identity_edit_v1_22026-09-030.81×932/2202/52/5
krea2_identity_edit_v1_22026-09-040.58×57/242/32/3
krea2_turbo_fp8_scaled2026-08-210.40×455/9372/62/6
krea2_turbo_fp8_scaled2026-08-220.30×474/13642/62/6
krea2_turbo_fp8_scaled2026-08-230.42×628/11312/62/6
krea2_turbo_fp8_scaled2026-08-240.58×60/1922/62/6
krea2_turbo_fp8_scaled2026-08-260.42×563/11242/62/6
krea2_turbo_fp8_scaled2026-08-270.31×434/12642/62/6
krea2_turbo_fp8_scaled2026-08-280.66×134/222/42/4
krea2_turbo_fp8_scaled2026-08-290.36×346/3552/62/6
krea2_turbo_fp8_scaled2026-08-300.57×30/402/32/3
krea2_turbo_fp8_scaled2026-08-310.52×789/572/52/5
krea2_turbo_fp8_scaled2026-09-010.44×502/482/32/3
krea2_turbo_fp8_scaled2026-09-020.58×29/162/52/5
krea2_turbo_fp8_scaled2026-09-031.32×671/1812/52/5
krea2_turbo_fp8_scaled2026-09-040.55×33/432/52/5
one_obsession_v22_fp162026-08-210.21×229/1812/62/6
one_obsession_v22_fp162026-08-220.20×244/4532/62/6
one_obsession_v22_fp162026-08-230.34×88/1922/62/6
one_obsession_v22_fp162026-08-240.35×101/1242/42/4
one_obsession_v22_fp162026-08-260.35×194/7342/62/6
one_obsession_v22_fp162026-08-270.41×186/4292/62/6
one_obsession_v22_fp162026-08-310.44×47/482/52/5
one_obsession_v22_fp162026-09-030.84×180/572/42/4
z_image_bf162026-08-210.51×12/152/42/4
z_image_bf162026-08-220.51×12/152/42/4
z_image_bf162026-08-230.45×39/332/62/6
z_image_bf162026-08-240.51×40/222/62/6
z_image_bf162026-08-260.46×44/1372/62/6
z_image_bf162026-08-270.70×12/432/52/5
z_image_bf162026-08-290.37×32/242/42/4
z_image_bf162026-09-030.67×12/182/42/4
z_image_turbo_bf162026-08-230.31×25/722/62/6
z_image_turbo_bf162026-08-240.28×32/572/42/4
z_image_turbo_bf162026-08-260.36×13/882/62/6
What this cannot isolate. A ratio compares two deployments - card, host, virtualisation layer and dispatch together. Rows marked † carry a known host confound and are excluded from headline ranges. Attempts are not independent observations; replication is counted in physical cards and card-days (methodology §7e).

Bring your own hourly quotes to /quote to turn any of these ratios into a break-even price for your workload. Raw cells: gpu-goodput-ratios.csv.