What's Changed

• Fix yolo26 XL Seg aliases by @Matvezy in #1919
• Fix rfdetr instance segmentation and object detection postprocessing error by @grzegorz-roboflow in #1920
• Add docs about YOLO26 by @PawelPeczek-Roboflow in #1923

Full Changelog: v0.64.2...v0.64.3

What's Changed

• Add more yolo26 aliases by @Matvezy in #1918

Full Changelog: v0.64.1...v0.64.2

What's Changed

• Add yolo26 aliases by @Matvezy in #1917

Full Changelog: v0.64.0...v0.64.1

💪 Added

YOLO26 in inference

Legendary effort of @probicheaux 🏅 and models team made it! Shortly after release, we have a support for new Ultralytics model - YOLO26. If you are interested - learn more about the model, or you can cut the line and jump straight into testing the model on Robolfow Platform.

🔎 How to use the new model?

• Train the model on your data using Robolfow Platform.
• Use it in Workflows or directly in inference

As easy as that 😀

# install inference - uv pip install inference or uv pip install inference-gpu
import cv2
import supervision as sv
from inference import get_model

model = get_model("<id-of-your-yolo26-model>", api_key="<your_api_key>")
image = cv2.imread("<your-image-path>")

predictions = sv.Detections.from_inference(model.infer(image)[0])

annotator = sv.BoxAnnotator()
annotated_image = annotator.annotate(image.copy(), predictions)

New VLMs 🤝 Workflows - Gemini 3 Flash and Claude Opus 4.5

Workflows ecosystem is constantly growing 🤯 Thanks to @yeldarby contributions, we are now able to expose new VLMs in Workflows - plus, for both Gemini and Claude models family, Roboflow now provides managed API keys - so Roboflow account alone is enough - no manual setup of your API keys.

🔧 Bug fixes

• Adds support for selectors in Union[List[...], Selector(...)] patterns in Workflows Execution Engine by @joaomarcoscrs in #1764

🚧 Maintanence

• Fix assertions in GPU tests and add fixture to clean-up memory by @PawelPeczek-Roboflow in #1904
• Fix assertions in GPU tests by @PawelPeczek-Roboflow in #1905
• add yolo and rfdetr search terms to obj-det and keypoint-det workflow blocks by @mkaic in #1906
• Fix documentation links by @sergii-bond in #1908

🥼 New inference - progress update

• Rename inference-exp -> inference-models by @PawelPeczek-Roboflow in #1851
• Fix missing changes of inference-exp to inference-models by @PawelPeczek-Roboflow in #1902
• Fix inference-models again by @PawelPeczek-Roboflow in #1903

🏅 New Contributors

• @sergii-bond made their first contribution in #1908

Full Changelog: v0.63.5...v0.64.0

💪 Added

• feature: Event Detection Log Block by @jeku46 in #1894
  We have new analytics block in workflows - Event Detection Log Block
  This block takes input from the bytetracker (or any other tracker) and stores a log of events. The events store the time an object (unique tracker id) was first seen, last seen, its tracker id, and class, like shown below:

{
  "1": {
    "tracker_id": 1,
    "class_name": "dog",
    "first_seen_frame": 1,
    "last_seen_frame": 107,
    "frame_count": 102,
    "first_seen_relative": 0.0,
    "last_seen_relative": 3.533333333333333,
    "first_seen_timestamp": 1765744769.0,
    "last_seen_timestamp": 1765744772.5333333
  }
}
Pending events (2):
{
  "5": {
    "tracker_id": 5,
    "class_name": "dog",
    "first_seen_frame": 87,
    "last_seen_frame": 87,
    "frame_count": 1,
    "first_seen_relative": 2.8666666666666667,
    "last_seen_relative": 2.8666666666666667,
    "first_seen_timestamp": 1765744771.8666666,
    "last_seen_timestamp": 1765744771.8666666
  },
  "6": {
    "tracker_id": 6,
    "class_name": "dog",
    "first_seen_frame": 93,
    "last_seen_frame": 105,
    "frame_count": 6,
    "first_seen_relative": 3.066666666666667,
    "last_seen_relative": 3.466666666666667,
    "first_seen_timestamp": 1765744772.0666666,
    "last_seen_timestamp": 1765744772.4666667
  }
}

With this block our users can now easily track lifecycle of detected and tracked objects!

• feature: Add Text Display Visualization Block by @rafel-roboflow in #1895
  We have new visualization block in workflows - Text Display Visualization Block, users can now add parametrized text element on annotated image, showing live results of other blocks and making it much easier to showcase real value brought by workflow!

🚧 Maintenance

• Add new RF-DETR-Seg sizes by @probicheaux in #1897
• fix: allow workflow input selectors for motion detection parameters by @sberan in #1898

📖 Documentation improvements

• Enhance documentation for visualizations workflow blocks by @patricknihranz in #1880
• Docs transformations by @patricknihranz in #1882
• Docs sinks by @patricknihranz in #1883
• Enhance documentation for analytics workflow blocks by @patricknihranz in #1884
• Docs cache by @patricknihranz in #1885
• Docs classical cv by @patricknihranz in #1886
• Enhance documentation for flow_control workflow blocks by @patricknihranz in #1887
• Docs formatters by @patricknihranz in #1888
• Docs fusion by @patricknihranz in #1889
• Docs math by @patricknihranz in #1890
• Docs sampling by @patricknihranz in #1891
• Enhance documentation for secrets_providers workflow blocks by @patricknihranz in #1892

🏅 New Contributors

• @jeku46 made their first contribution in #1894
• @rafel-roboflow made their first contribution in #1895
• @patricknihranz made their first contribution in #1880

Full Changelog: v0.63.4...v0.63.5

💪 Added

• Thanks to @mkaic , Workflows support DepthAnythingV3 🥳

• Thanks to @Matvezy , Workflows support Qwen3vl 🥳

• Workflows now make it easier to transform detections results between execution stages through use of new detections rollup block, change by @lou-roboflow in #1853

• API Key Passthrough for Gemini was added to workflows, change also adds support for newer Gemini models (Gemini 3 Pro, 2.5 Pro/Flash/Flash-Lite, 2.0 Flash/Flash-Lite) and thinking_level parameter for Gemini 3+ models - added by @yeldarby in #1869

🚧 Maintenance

• Fix gaze_test by @grzegorz-roboflow in #1870
• Remove non-existing link to foundation/easy_ocr.md by @mattmazzola in #1873
• Fix watchdog on_timeout by @grzegorz-roboflow in #1875
• Remove extra detections_rollup folder by @lou-roboflow in #1876
• Update Qwen to Flash Attn and Fix Transformers Bug by @Matvezy in #1878
• Preserve class names for rolled up detections by @lou-roboflow in #1877
• Manually override wrong task type in RFDETRInstanceSegmentation by @mkaic in #1852
• timm 1.0.23 introduces error: … by @grzegorz-roboflow in #1879
• Fix paligemma sdpa issue by @Matvezy in #1893

New Contributors

• @mattmazzola made their first contribution in #1873

Full Changelog: v0.63.3...v0.63.4

💪 Added

• Add Managed API Key for Twilio SMS/MSS Block by @yeldarby in #1860
• Implement v2 Camera Focus block with Tenengrad measure and improved visualization by @shntu in #1857
• Add mjpeg support to webrtc worker by @grzegorz-roboflow in #1867
• Implement ACK-based flow control for Python SDK WebRTC batch processing by @digaobarbosa in #1865

🚧 Maintenance

• Fix line counter visualization line opacity by @grzegorz-roboflow in #1868
• ModalConnection closed when running on modal through TURN server by @digaobarbosa in #1858
• Added offline weights download documentation by @renzon in #1847
• Add sam3d disclaimer by @Matvezy in #1854
• Fix small bug in motion detection by @lou-roboflow in #1850
• Add stop delay to continue if by @lou-roboflow in #1844

🥼 inference-exp improvements

• Fix/dinov3 gpu integration tests by @grzegorz-roboflow in #1849

New Contributors

• @renzon made their first contribution in #1847

Full Changelog: v0.63.2...v0.63.3

What's Changed

• SAM2 in inference-exp by @PawelPeczek-Roboflow in #1841
• Add README covering webrtc streaming by @grzegorz-roboflow in #1842
• Sam3d by @Matvezy in #1784
• Fixresnet typo by @Matvezy in #1846
• Fix issue with OWLv2 serialisation and deserialisation by @PawelPeczek-Roboflow in #1848

Full Changelog: v0.63.1...v0.63.2

💪 Added

GPT-5.2 in Workflows 🧠

🕵️ Motion Detection block in Workflows

ResNet weights for classifiers

Thanks to @Matvezy, we have good old ResNet classifiers registered under clear aliases:

from inference import get_model

model = get_model("resent18", api_key="<YOUR_API_KEY>")

# Check out other models: "resent34", "resent50", "resent101"

🚧 Maintenance

• add missing extra-require by @japrescott in #1806
• Add docs for Jetson 6.2 base image ENT-645 by @sberan in #1803
• removing uploading frame through datachannel by @digaobarbosa in #1829
• Do not trim API key hash by @grzegorz-roboflow in #1831
• Pass is_preview when testing workflows by @grzegorz-roboflow in #1830
• Weights for Seg Preview by @Matvezy in #1804
• Bump version to 0.63.0 in prep for release by @PawelPeczek-Roboflow in #1838
• Fix detection metadata length mismatch in workflows by @tonylampada in #1835

🥼 inference-exp improvements

• Add OWLv2 and RF instant models to inference-exp by @PawelPeczek-Roboflow in #1807
• Loose assertions in RF instant model e2e test by @PawelPeczek-Roboflow in #1809
• inference-exp 0.17.2 by @PawelPeczek-Roboflow in #1811
• add dinov3 to inference experimental by @grzegorz-roboflow in #1810
• SAM in inference-exp by @PawelPeczek-Roboflow in #1833
• DinoV3 classification - torch backend by @grzegorz-roboflow in #1837
• Import inference-exp within stream_url_to_cache by @grzegorz-roboflow in #1839

Full Changelog: v0.62.5...v0.63.1

💪 Added

• Streaming for vlms by @Matvezy in #1795
• Add SAM3 v3 block with RLE mask support by @lrosemberg in #1801
• Add changes to ensure jetson environment is specified correctly by @PawelPeczek-Roboflow in #1796
• webrtc sdk addding new parameters to region and plans by @digaobarbosa in #1797

🔧 Fixed

• Fix workflow builder Private Network Access for Chrome 142+ by @alexnorell in #1794
• removing timeout for video frames by @digaobarbosa in #1798
• Fix for video tracks with video files by @digaobarbosa in #1799

Full Changelog: v0.62.4...v0.62.5