Compare D-FINE vs YOLO26 live
Run the same image across every model that supports a task and compare their outputs side-by-side.
These models don't share enough common tasks for a side-by-side demo. See the comparison table below for their capabilities.
D-FINE vs YOLO26 Comparison Table
| Property | D-FINE | YOLO26 |
|---|---|---|
| Organization | USTC | Ultralytics |
| Category | open | open |
| Modality | vision | vision |
| Release Date | Oct 2024 | Oct 2025 |
| Context Window | — | — |
| Parameters | 4M-62M | 2.4M-55.7M |
| License | Apache 2.0 | AGPL 3.0 |
| Commercial use | free for commercial use | commercial use needs a license (included on Roboflow cloud) |
| COCO Benchmark object detection, mAP 50:95 on COCO val2017 | ||
| Best size | Not benchmarked | 56.3Extra Large · 55.7M params |
| Smallest size | Not benchmarked | 39.9Nano · 2.4M params |
| mAP range | Not benchmarked | 39.9–56.3across 5 sizes |
| Model Sizes COCO mAP 50:95, tested resolution and params per size | ||
| Nano | Not available | 39.9640×640 · 2.4M params |
| Small | Not available | 47.2640×640 · 9.5M params |
| Medium | Not available | 52.0640×640 · 20.4M params |
| Large | Not available | 53.6640×640 · 24.8M params |
| XL | Not available | 56.3640×640 · 55.7M params |
| COCO val2017 mAP 50:95 and parameter counts from leaderboard.roboflow.com. Size names don't line up across model families, so compare sizes with similar parameter counts. | ||
| Vision Tasks | ||
| Object Detection | Supported | Demo (COCO) |
| Instance Segmentation | Not listed | Demo (COCO) |
| Model Features | ||
| Real-Time Vision | Supported | Not listed |
D-FINE vs YOLO26: Overview
D-FINE is a real-time object detection model introduced in October 2024 by researchers at the University of Science and Technology of China. It builds on the DETR family of transformer-based detectors by reformulating bounding box regression as a Fine-grained Distribution Refinement task. Rather than predicting box coordinates directly, D-FINE iteratively refines probability distributions over coordinate offsets across decoder layers, which provides finer localization granularity without adding inference cost. The architecture also replaces the encoder's CSP blocks with GELAN modules and inserts a Target Gating Layer after the decoder's cross-attention to reduce representational entanglement across queries. A second contribution, Global Optimal Localization Self-Distillation, transfers localization knowledge from refined deeper-layer predictions back to earlier decoder layers through internal self-distillation.
D-FINE is released in five model sizes (Nano, Small, Medium, Large, and X), with D-FINE-L achieving 54.0% AP on the Microsoft COCO benchmark at 124 FPS on an NVIDIA T4 GPU, and D-FINE-X reaching 55.8% AP at 78 FPS. Pretraining on the Objects365 dataset further improves accuracy to 57.1% AP for the L variant and 59.3% AP for the X variant. The paper was accepted at ICLR 2025 as a Spotlight. Code and pretrained weights are released under the Apache 2.0 license, making the model suitable for commercial use.
YOLO26 is a real-time object detection model developed by Ultralytics, released in October 2025. It introduces a native end-to-end, NMS-free architecture that eliminates the Non-Maximum Suppression post-processing step, reducing CPU latency by up to 43% for the Nano variant compared to NMS-dependent versions. The model incorporates the MuSGD optimizer and ProgLoss with STAL for improved training stability and small-object detection, and removes Distribution Focal Loss to ensure maximum compatibility with ONNX and TensorRT export targets.
YOLO26 supports object detection, instance segmentation, pose estimation, and oriented bounding box detection within a unified framework, with model sizes available from Nano to Extra Large. Its NMS-free design makes it particularly well suited for deployment scenarios where post-processing overhead is a bottleneck, such as embedded systems and real-time edge inference pipelines.
