gNucleus AI

There's understandable concern about AI replacing engineers. History suggests something different. CAD didn't replace engineers. Simulation didn't replace engineers. Cloud PLM didn't replace engineers. Each technology expanded what engineers could accomplish. Engineering AI is likely to do the same—automating repetitive work so engineers can focus on innovation, trade-offs, and complex decisions. The role evolves. The need for engineering expertise doesn't disappear. #Engineering #AI #Innovation @gNucleusAI

🎉 Congratulations to Ryan Marten, Alex Shaw, Andy Konwinski, Ludwig Schmidt, the @harbor team, Laude Institute, and the entire community on the official launch of Frontier-Bench! 🚀We are incredibly proud that gNucleus tasks are officially featured as part of the launch dataset, powering the CAD evaluation track! ⚙️ ⚙️As a data partner for Frontier-Bench alongside incredible teams like tge sponsors: @modal_labs, Anthropic, OpenAI, Google , and data partners ScaleAI, SnorkelAI, Openbenchmarks, Turing, Boolean AI, and Handshake, we loved collaborating to build the next generation of AI benchmarks! Huge thanks to the project leaders & contributors for pulling this milestone together! #SiliconValley #gNucleus AI #gNucleusAI #Innovation #Leadership

@ryan_marten @modal_labs @AnthropicAI @OpenAI @Google @scale_AI @SnorkelAI @turingcom @joinHandshake Proud to be a data partner for Frontier-Bench alongside incredible teams!

Frontier-Bench was made possible by our sponsors. Thank you to our compute sponsors @modal_labs, @AnthropicAI, @OpenAI, @Google and our data partners @scale_AI, @SnorkelAI Open Benchmarks, @turingcom, @gNucleusAI, Boolean AI, and @joinHandshake. Frontier-Bench is hosted by @harborframework and @LaudeInstitute.

We’re releasing Frontier-Bench: a benchmark that measures and evolves with the frontier of agent work. Built by the team behind Terminal-Bench and Harbor, Frontier-Bench is an on-going community effort. Frontier-Bench v0.1 contains 74 tasks on which the best agents score ~34%

Proud to be a data partner for Frontier-Bench @frontierbench alongside incredible teams like Bench's Sponsors: @modal_labs, @AnthropicAI, @OpenAI, @Google Bench's data partners @scale_AI, @SnorkelAI Open Benchmarks, @turingcom, @gNucleusAI, Boolean AI, and @joinHandshake! Testing frontier models on complex, real-world CAD and physical engineering data is essential for the next generation of AI. Huge thanks to @ryan_marten, @alexgshaw, other project leaders & contributors, @harborframework and @LaudeInstitute for pulling this milestone together! 🚀 x.com/ryanmart3n/sta…...

Check out the featured task and other tasks (part of Frontier Bench @frontierbench ) x.com/gNucleusAI/sta…

🚀 Huge milestone for gNucleus and the entire generative engineering space! As a contributor to Frontier-Bench, we are excited to have our CAD evaluation tasks was featured in Anthropic’s announcement of Claude Opus 5. As our CTO Mei Chen puts it: Creating meaningful evaluations for AI in engineering is difficult. The tasks must reflect the geometric reasoning, tool use, iteration, and real-world friction involved in professional CAD workflows. Congrats to Ryan Marten Alex Shaw and Anthropic team! @claudeai Check out our post for the full backstory! 👇 linkedin.com/feed/update/ur… #IndustrialAI #CAD #LLM #AI #gNucleus AI

@alexgshaw Congrats to Ryan Marten Alex Shaw

Benchmark your agent's ability to do... anything! Frontier-Bench is available today in the Harbor CLI, on Harbor Hub, and on GitHub. Frontier-Bench is also our way of demonstrating how benchmarks should be continuously developed, versioned, and maintained, and how prior results can be migrated alongside benchmark updates.

@ryan_marten Congrats to Ryan Marten Alex Shaw

Read the Claude Opus 5 announcement: anthropic.com/news/claude-op… Learn more about Frontier-Bench: frontierbench.ai Explore our FreeCAD tasks in Frontier-Bench v0.1: hub.harborframework.com/datasets/front… #ClaudeOpus5 #Anthropic #FrontierBench #AIforCAD #FreeCAD #EngineeringAI #GenerativeAI #Manufacturing #gNucleusAI

The featured task is : frontier-bench/freecad-platform-drawing It is one of three complex FreeCAD evaluations contributed by gNucleus AI to Frontier-Bench v0.1: • freecad-platform-drawing — an image-to-CAD reconstruction challenge • freecad-impeller — a complex text-to-CAD modeling task • freecad-spring-clip — another rigorous text-to-CAD evaluation @claudeai @alexgshaw @terminalbench @harborframework

A proud moment for the gNucleus AI team: one of our CAD evaluation tasks was featured in Anthropic’s announcement of Claude Opus 5. 🚀

Introducing Claude Opus 5. It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.

@claudeai 🚀As a contributor to Frontier-Bench, we are excited to have our CAD evaluation tasks was featured in announcement of Claude Opus 5. x.com/gNucleusAI/sta…

On several coding and knowledge work evaluations, Opus 5 is the new state-of-the-art:

⚡ Huge thanks to Michael Finocchiaro for the fantastic spotlight and thoughtful conversation! You nailed our core thesis: building true engineering AI isn't about wrapping existing LLMs, it's about building models that natively understand complex CAD and production-grade designs to embed into workflows across the whole lifecycle. ✅ Coming from an industry veteran who tracks over 750 Industrial AI and PLM startups, this is massive validation for our team. We’re also so grateful to see industry leaders and pioneers resonating with this vision! 🙌 Jalpan Dave Hunter Horne Manoj Krishnan Michael Wakefield Ken Coulter Florin Anesia Brion Carroll (II) Jey Michaelraj Check out full breakdown below! 👇 linkedin.com/feed/update/ur…

To all the newcomers excited to try Opus 4.8-level models at home: welcome to OpenWeightLand! Things work a little differently here than in ClosedSourcistan. Might seem strange at first but you'll quickly get used to it: - there are many providers for the same model and they compete on price and features. - as a result intelligence is abundant and typically much cheaper - you can run the model on-prem, in your region, locally, or with the provider of your choice - you can fine-tune it, modify it, and build businesses on top of it without asking anyone for permission Turns out open weights create markets, not kingdoms. A good central train station to start exploring is the Hugging Face page for GLM-5.2 under "Use this model": -> huggingface.co/zai-org/GLM-5.2 And if you just want to chat with it, it's free on HuggingChat: -> huggingface.co/chat/

Congrats to @MikushRab on the launch of CADGenBench! 🚀🏗️ Make sure to check out both CADGenBench for evaluation, and our very own Parametric CAD Bench x.com/gNucleusAI/sta… on Hugging Face - packed with 10k+ high-fidelity B-Rep samples to train models that actually understand engineering logic. If you are working on the future of generative engineering, you now have two incredible resources to benchmark your models. Let's push the boundaries of physical AI together! 👇 x.com/gNucleusAI/sta… #OpenSource #GenerativeAI #CAD #MachineLearning @huggingface

🌐 Let’s Collaborate:We are actively looking to expand our footprint with academic institutions, innovation hubs, and professional communities. Whether you want to collaborate on our recent bench launch, stress-test our upcoming roadmaps, or bring a live tech case study to your classroom - the sky is the limit. If you are a professor, community leader, or ecosystem partner who wants to build something together, send me a DM or reach out to our team! 🚀 #Operations #ExecutiveLeadership #OpenInnovation #EMBA #gNucleus #Partnerships

🔗 Read the full Bench Report: dev.cadbench.ai/news/parametri…