Skip to content
ModelsThatRun

All models

32 model families. Latency, VRAM and accuracy are measured on our RTX 4090; the paper's accuracy sits under ours, labelled.

Object detection

ModelAccuracy (paper under it)4090 latency (input under it)VRAM peakLicense
D-FINE5 sizes: n, s, m, l, x42.6 box AP42.8 COCO AP val paper21.9 ms640x640563 MiBCommercial OK
DEIMv24 sizes: s, m, l, x50.6 box AP50.9 COCO AP paper27.4 ms640x640575 MiBNon-commercial
Grounding DINO2 sizes: tiny, base48.3 box AP48.4 COCO AP, zero-shot paper118 ms1199x8002.9 GiBCommercial OK
GTR4 sizes: s, m, l, x53.3 box AP53.6 COCO AP paper34.0 ms640x640636 MiBCommercial OK
LLMDet3 sizes: tiny, base, large55.3 box AP164 ms1199x8003.0 GiBCommercial OK
MM-Grounding-DINO3 sizes: tiny, base, large50.3 box AP50.6 COCO mAP, zero-shot paper164 ms1199x8003.0 GiBCommercial OK
RF-DETR4 sizes: nano, small, medium, large47.9 box AP48.4 COCO AP50:95 paper9.8 ms384x384803 MiBCommercial OK
RT-DETRv44 sizes: s, m, l, x49.6 box AP49.8 COCO AP paper24.5 ms640x640639 MiBCommercial OK
YOLO115 sizes: n, s, m, l, x39.1 box AP39.5 COCO mAP50-95 paper14.0 ms640x448572 MiBConditional
YOLO265 sizes: n, s, m, l, x40.5 box AP40.9 COCO mAP50-95 paper14.2 ms640x448571 MiBConditional
YOLOv85 sizes: n, s, m, l, x37.0 box AP37.3 COCO mAP50-95 paper11.6 ms640x448514 MiBConditional

Instance segmentation

ModelAccuracy (paper under it)4090 latency (input under it)VRAM peakLicense
EoMT2 sizes: large-640, large-128044.8 mask AP45.2 COCO mAP (mask) paper44.7 ms640x6406.4 GiBCommercial OK
GTR-Seg4 sizes: s, m, l, x44.9 mask AP45 COCO mask AP paper35.6 ms640x640677 MiBCommercial OK
RF-DETR-Seg6 sizes: nano, small, medium, large, xlarge, 2xlarge40.1 mask AP40.3 COCO AP50:95 (mask) paper12.5 ms312x312887 MiBCommercial OK
YOLO11-seg5 sizes: n, s, m, l, x31.8 mask AP32 COCO mAPmask 50-95 paper14.5 ms640x448596 MiBConditional
YOLO26-seg5 sizes: n, s, m, l, x33.5 mask AP33.9 COCO mAPmask 50-95 (e2e) paper15.5 ms640x448579 MiBConditional
YOLOv8-seg5 sizes: n, s, m, l, x30.1 mask AP30.5 COCO mAPmask 50-95 paper12.8 ms640x448590 MiBConditional

Image matting

ModelAccuracy (paper under it)4090 latency (input under it)VRAM peakLicense
BEN21 size: baseNot yet scored112 ms1024x10243.6 GiBCommercial OK
BiRefNet3 sizes: general, hr, liteNot yet scored92.8 ms1024x10243.5 GiBCommercial OK

Depth estimation

ModelAccuracy (paper under it)4090 latency (input under it)VRAM peakLicense
Depth Anything 3 Metric1 size: largeNot yet scored26.9 ms504x3362.1 GiBCommercial OK
Depth Anything 3 Mono1 size: largeNot yet scored27.3 ms504x3362.1 GiBCommercial OK
Depth Anything V23 sizes: small, base, largeNot yet scored17.1 ms770x518858 MiBCommercial OK smallNon-commercial base, large
MoGe-23 sizes: vits, vitb, vitlNot yet scored44.1 ms1022x6861.9 GiBCommercial OK

Pose estimation

ModelAccuracy (paper under it)4090 latency (input under it)VRAM peakLicense
GTR-Pose4 sizes: s, m, l, xNot yet scored70.1 COCO keypoint AP paper34.0 ms640x640639 MiBCommercial OK
RF-DETR Keypoint1 size: previewNot yet scored71.8 COCO keypoint AP50:95 paper16.7 ms576x5761.3 GiBCommercial OK
YOLO11-pose5 sizes: n, s, m, l, xNot yet scored50 COCO mAPpose 50-95 paper14.2 ms640x448574 MiBConditional
YOLO26-pose5 sizes: n, s, m, l, xNot yet scored57.2 COCO mAPpose 50-95 (e2e) paper14.9 ms640x448580 MiBConditional
YOLOv8-pose5 sizes: n, s, m, l, xNot yet scored50.4 COCO mAPpose 50-95 paper12.6 ms640x448514 MiBConditional

Chat

ModelAccuracy (paper under it)4090 decodeVRAM peakLicense
Gemma 42 sizes: 12b-q4_0, 31b-q4_089.3 IFEval97.4 tok/s7.7 GiBCommercial OK
Qwen32 sizes: 0.6b-q8_0, 8b-q4_k_m57.3 IFEval54.5 IFEval strict prompt paper533 tok/s1.9 GiBCommercial OK
Qwen3.51 size: 9b-q4_k_m83.5 IFEval91.5 IFEval (thinking mode not stated) paper134 tok/s5.6 GiBCommercial OK
Qwen3.81 size: 27b-q4_k_m81.0 IFEval43.0 tok/s18.3 GiBCommercial OK