{"url":"https://newsforagent.com/news/2026-11-18/xailient-nist-border-kiosk-accuracy","markdown_url":"https://newsforagent.com/news/2026-11-18/xailient-nist-border-kiosk-accuracy.md","title":"Xailient Tops NIST Border-Kiosk Test With 95.4% Accuracy","headline":"Xailient's xailient_001 Model Leads NIST Border-Kiosk Test Among 106 Fast Developers","description":"Xailient's xailient_001 model matched 95.4% of genuine Border-Kiosk pairs at a 1-in-100,000 false match rate.","section":"Product Launches","published_at":"2026-11-18T03:23:03+00:00","modified_at":"2026-11-18T03:23:03+00:00","author":"News Desk","key_takeaways":["Xailient's xailient_001 model ranked first among 106 developers with template generation of 500 milliseconds or less in NIST's Border-Kiosk test.","The model matched 95.4% of genuine image pairs at a false match rate of 1 in 100,000, using 901 MB of memory and 497 milliseconds of template time.","Xailient's production system evaluates multiple frames per subject, and ISO live-subject testing matched 250 moving subjects against a gallery of 10,000 candidates with 99.5% recognition and zero false matches."],"companies":[{"name":"Xailient","is_primary":true,"profile_url":"https://akta.pro/company-profiles/00028n8/xailient","website":"https://www.xailient.com"},{"name":"NIST","is_primary":false,"profile_url":"https://akta.pro/company-profiles/0003lcn/nist","website":"http://www.nist.gov"}],"source":{"publisher":"PR Newswire","title":"Xailient Ranks #1 for Face Recognition Accuracy in Real-world On-Device AI","url":"https://www.prnewswire.com/news-releases/xailient-ranks-1-for-face-recognition-accuracy-in-real-world-on-device-ai-302943106.html"},"other_sources":[{"publisher":"Bloomberg","url":"https://www.bloomberg.com/news/newsletters/2026-11-15/apple-siri-ai-app-integrations-power-smart-glasses-iphone-duo-2-face-id-plan"},{"publisher":"Techday","url":"https://techday.co.uk/story/gbg-adds-realeyes-face-verification-to-go-platform"},{"publisher":"Sourcesecurity","url":"https://www.sourcesecurity.com/news/suprema-ai-vision-facial-authentication-sgpp-co-2366-ga.1794377199.html?ref=rss-an"},{"publisher":"Biometric Update","url":"https://www.biometricupdate.com/202611/sigur-integrates-iris-id-biometrics-into-centralized-access-management"}],"tags":["Product Launches","Xailient","NIST"],"body_markdown":"[Xailient](https://akta.pro/company-profiles/00028n8/xailient) said its xailient_001 face recognition model achieved the highest accuracy in the U.S. National Institute of Standards and Technology's Face Recognition Technology Evaluation Border-Kiosk test among 106 developers whose template generation time is 500 milliseconds or less. The Las Vegas announcement, dated November 17, 2026, was issued through [PR Newswire](https://www.prnewswire.com/news-releases/xailient-ranks-1-for-face-recognition-accuracy-in-real-world-on-device-ai-302943106.html).\n\n## How xailient_001 Performed In Border-Kiosk\n\nAt a false match rate of 1 in 100,000, xailient_001 correctly matched 95.4% of genuine Border-Kiosk image pairs. At the stricter rate of 1 in 1,000,000, the model recorded the equivalent of 94.0% of genuine pairs matched. On the easier Border-Border test, it reached a 0.55% false non-match rate at a false match rate of 1 in 1,000,000, or approximately 99.45% genuine-match accuracy.\n\n[NIST](https://akta.pro/company-profiles/0003lcn/nist) tested the same model Xailient deploys on edge hardware with roughly 1 TOPS of NPU compute. The model generated face templates in 497 milliseconds and used 901 MB of memory plus 280 MB of configuration data, the lowest figures in Xailient's comparison of six major competitors. Xailient said the nearest competitor required 1.8 times more memory and 1.6 times more configuration data.\n\nNIST's Border-Kiosk test is the only static-image 1:1 test involving non-cooperative subjects, with a touchscreen below the camera producing downward facial pitch that can exceed 45 degrees. Kiosk images are compared with border-crossing images, matching two unposed photos. \"Face recognition performance in the real world is not just about achieving the best score with unlimited compute,\" said Lars Oleson, CEO of Xailient. \"It is about delivering high accuracy on the hardware customers actually deploy.\"\n\n## Xailient's Edge And Casino Partnerships\n\nAccording to akta.pro [company records](https://akta.pro/company-data), Xailient was founded in 2019, is headquartered in Sunnyvale, and has raised $3.4 million in total funding across three rounds, the last a $2.2 million seed in June 2021. The same records list a partnership with Sony Semiconductor Solutions to offer face recognition AI for the IMX500 intelligent vision sensor, achieving 97.8% accuracy up to 3 meters.\n\nThe records also list SYNK Vision, launched with Konami Gaming in July 2024 to replace magnetic player tracking cards with facial recognition on casino slot machines, and a January 2026 partnership with Durin as one of five technology providers for the MagicKey home-entry system combining UWB proximity, on-device facial recognition and voice biometrics.\n\n## Face Verification Deals Across The Sector\n\nGBG integrated Realeyes face verification into its Go identity platform, using a three-second selfie to return a reusable face credential, [Techday](https://techday.co.uk/story/gbg-adds-realeyes-face-verification-to-go-platform) reported on November 11, 2026. Sigur integrated Iris ID's iris and facial recognition terminals into its access management platform in software version 1.8.5.69 and later, [Biometric Update](https://www.biometricupdate.com/202611/sigur-integrates-iris-id-biometrics-into-centralized-access-management) reported the same day.\n\nSuprema planned its 2026 global partner conference in Gyeongju, South Korea, for November 11–12, highlighting facial authentication server matching at a 1:2 million scale, [Sourcesecurity](https://www.sourcesecurity.com/news/suprema-ai-vision-facial-authentication-sgpp-co-2366-ga.1794377199.html?ref=rss-an) reported on November 11, 2026.\n\n## Xailient's Live-Subject Testing\n\nXailient said the NIST result follows independent ISO live-subject testing in which 250 diverse, moving subjects were matched against a gallery of 10,000 registered candidates in real time, with 99.5% successful recognition and zero false matches. \"By matching on the camera, edge deployment avoids collecting bystander data - only recognizing opt-in and officially registered subjects,\" said Mallika Patel, VP of Product.\n","publisher":{"name":"News for Agent","url":"https://newsforagent.com"}}