Today we're introducing Precision AI Open Release · Vol. 1 — the first in a new series of open agricultural AI models, benchmarks, and tools built for the real-world. Most people know Precision AI for precision spraying and drones. But at our core, we're an agricultural AI company with a world-class applied research team — and today, we're opening up part of that work. We're starting with a strong result. AgriEmbedding ranks first on both AgriStress-500 and a public agricultural image retrieval benchmark. It outperforms every DINOv2 and TIPSv2 model we evaluated, while using roughly 4× fewer parameters and half the embedding dimensions of the largest top-performing model in the comparison. Vol. 1 includes: • AgriEmbedding API: free, hosted, agriculture-specific image embeddings for retrieval, classification, and spatial vision. Both dense and compact options. • AgriStress-500: a benchmark built from 500 challenging images captured in real fields. • precisionai-agrieval — a Python library for evaluating agricultural vision models. These are the first open components of our agricultural AI research stack, spanning self-supervised learning, real-world model evaluation, and emerging multimodal and agentic workflows. We're inviting AI researchers, agtech engineers, robotics teams, agronomists, and academic labs to use the tools. Try it. Test it. Build with it. 🔗 AgriEmbedding API: embeddings.precision.ai 🔗 AgriStress-500 on HuggingFace: lnkd.in/gjCuq7MX 🔗 precisionai-agrieval: lnkd.in/geasrh58 Vol. 1 is just the beginning. We'll be releasing more agricultural AI research and tools on a regular cadence.

Every AgriEmbeddings API response now includes an Out-of-Distribution score. The embedding shows how the model represents an input. The OOD score shows how far that input sits from the model’s training distribution. That matters in agriculture, where crops, growth stages, geography, cameras, lighting, and seasons constantly shift the data. Use OOD scoring to flag unfamiliar imagery, filter datasets, monitor drift, and set reliability thresholds for downstream workflows. The threshold is yours to define based on the application. Live now in AgriEmbeddings: embeddings.precision.ai/

Evaluating a vision encoder for agricultural imagery? Read the protocol before you trust the leaderboard. We're opening our agricultural AI research to the community, and kicking it off with a technical evaluation report for our vision embedding model. Spraying a whole field to kill a few weeds wastes chemicals and speeds up herbicide resistance. Smarter treatment means identifying plants right down to the species, which is exactly where resistance management happens. This report puts the model to the test on how well it separates crops, weeds, and soil all the way to the species level. We published the full measurement method and scored the model on both whole-image retrieval and dense local prediction, at coarse and species granularity, including the tasks where baselines came out ahead. The benchmark images and labels are open, so you can run the numbers yourself. Read the report: research.precision.ai/seed_e…

Thank you! Our first open release hit #1 in agriculture on Hugging Face with thousands of AgriStress downloads. 🙏 Next up: AgriEmb, examples and reusable code that make our AgriEmbeddings API easy to use. More drops this week. 👉 github.com/Precision-AI-Inc/…

In just 24 hours, AgriStress-500 has achieved over 1,000 downloads and has become the #1 trending dataset on HuggingFace for agriculture. Thank you to everyone for your interest and support. It motivates us to release more tools in the coming weeks. Stay tuned!

Try the API. Stress the benchmark. Share the failure cases. Vol. 1 is the beginning. In Vol. 2, we’ll expand into out-of-distribution detection and introduce PAI Explorer: a dashboard for inspecting datasets, embeddings, model behavior, and task complexity. Our goal is to make agricultural AI easier to test, understand, and improve, while helping researchers and practitioners build useful applications for real field problems. If you’re working on AgAI, join the community, test what we’re releasing, and build with us.

In addition to our AgriEmbedding API, our model weights for fine-tuning are available on request under an academic license, a research or commercial partnership agreement. Reach out with your use case and we will respond within a week

Thank you for the overwhelming support we have received while launching the #StratusAirSprayer™. Your questions and feedback have been instrumental in making this groundbreaking product a reality. precision.ai/technology #PrecisionAI

Setup? A breeze! Be in the air in less than 15 minutes. The canopy is stored in an easy-to-carry chute bag, and the entire unit can be transported easily on a 12ft trailer. #StratusAirSprayer #PrecisionAI precision.ai/technology

The solution for low labor, sustainable, cost-efficient chemical application on the farm has been hindered by small drones that require battery changes every 4-6 minutes. By leveraging the power of a parachute, our AirSprayers™ handle payloads beyond traditional UAV limitations.

The #StratusAirSprayer™ will have onboard AI and computer vision for near-instant intelligent application, reducing environmental impact and ensuring chemical stays where it is needed. #PrecisionAI precision.ai/technology

The #StratusAirSprayer™ operates at less than $2.85/ac – a fraction of the cost of other solutions, ranging from $6 to $13/ac! How did we calculate that? We took the boom width, trip spray time, coverage rate (ac/hr) and # of flights needed per day compared to competitors.

Precision AI - Autonomous crop protection from the sky using artificial intelligence for precise application 🚀🌱

Our engineers have been committed to developing a solution that solves challenges faced by farmers. I’m proud of our team’s dedication, and can’t wait to see the impact the Stratus AirSprayer™ will have on your operations. - Andrew Streett, VP of Engineering

Why a parachute? Is it hard to setup? Can it do broadcast and spot spray application? What wind speeds can it handle? Answer these #StratusAirsprayer questions and more here: precision.ai/resources/1347/…

"The #StratusAirSprayer meets the rigorous demands of modern ag. Farmers have always driven us, and we're proud to deliver a product addressing the challenges of traditional UAVs. Thanks to our team and loyal customers for making this possible!" @dan_precisionai CEO

Finally, we can share some of the testing shots of our Gen 1 prototype of the Stratus AirSprayer™ from @Spaceport_NM and shot by the talented team at Lemon Light Productions 🛸

On Monday, we announced our new UAV design using the power of a parachute! Read about it and the Stratus AirSprayer's key benefits here: precision.ai/resources/1348/… #precisionai #StratusAirSprayer

A huge step forward for UAVs in #CanadianAg!