orq.ai

Alibaba Cloud's Model Studio Europe × Orq.ai 🇪🇺🚀 Model Studio is live in Frankfurt. Among the first European partners, Orq.ai integrates Qwen to deliver frontier AI performance with enterprise governance. Sovereign & GDPR-ready. #AlibabaCloud #OrqAI #Qwen #AInnovation #SovereignAI #Frankfurt #GenAIOps

𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝟰.𝟭 𝗵𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁: 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁 𝗜𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁𝘀 New features in the 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝘀 module alert! - 𝗔𝗱𝗱 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗼𝗿𝘀 𝗱𝘆𝗻𝗮𝗺𝗶𝗰𝗮𝗹𝗹𝘆  - 𝗘𝗻𝗮𝗯𝗹𝗲 𝗵𝘂𝗺𝗮𝗻 𝗿𝗲𝘃𝗶𝗲𝘄  - 𝗥𝗲𝗿𝘂𝗻 𝗲𝗿𝗿𝗼𝗿 𝗰𝗲𝗹𝗹𝘀

𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝟰.𝟭 𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗵𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁: 𝗠𝗲𝗺𝗼𝗿𝘆 𝗦𝘁𝗼𝗿𝗲𝘀 𝗠𝗲𝗺𝗼𝗿𝘆 𝗦𝘁𝗼𝗿𝗲𝘀 are persistent storage for agent memories, allowing your agents to maintain context and recall information across conversations

𝗥𝗲𝗹𝗲𝗮𝘀𝗲 𝟰.𝟭 𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗵𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁: 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗼𝗿𝗾 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗼𝗿𝗾 is a Python evaluation framework designed for running multiple experiments in parallel and measure AI performance directly from code. Google Colab link in the first comment

🚀 Orq.ai 4.1 is here Feature highlights dropping throughout the week: 🐍 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗼𝗿𝗾 𝗦𝗗𝗞 ⌨️ 𝗖𝗼𝗺𝗺𝗮𝗻𝗱 𝗕𝗮𝗿 🧪 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁 𝗜𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁𝘀 🛠️ 𝗧𝗼𝗼𝗹 𝗖𝗮𝗹𝗹 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝘀 🧠 𝗠𝗲𝗺𝗼𝗿𝘆 𝗦𝘁𝗼𝗿𝗲 𝗦𝘁𝘂𝗱𝗶𝗼

Las Vegas, here we come! 🎉 We will be exhibiting at @CES on 📅 January 6-9, 2026. See you at our📍 Booth #62100. Come watch live demos, meet our team, and discover how we're helping enterprises ship production-ready AI agents faster than ever. Who's joining us?

Ready to kick off 2026 ? Co-organized with miyagami.com, @lleverageai , and Adami.ai. We have prepared an amazing lineup to discuss the GenAI forecasts for 2026. Let’s kick off 2026 together! Sign up using Luma link in the comments!

What if you could ask multiple LLMs the same question and let them “debate” ? That’s the idea behind 𝗟𝗟𝗠 𝗖𝗼𝘂𝗻𝗰𝗶𝗹, proposed by Andrej Karpathy. Check out a short demo showing how this concept comes to life using Orq.ai’s AI Gateway.

By building a smart dual-model labeling pipeline, we generated 8,000 high-quality training samples for just $26, proving that industrial-grade fine-tuning is no longer reserved for those with massive budgets.

Even more impressive? With advanced prompt engineering, we unlocked +34% accuracy gains—equivalent to scaling models by 5–10× in size.

Can a 14B model really compete with a 100B+ model? Our latest research says: yes, when done right. orq.ai/blog/can-a-14b…

Thanks to our cracked 996 team, supportive clients, and investors. @AlexKoelpin @nanark_ @adriaanmol @galionexe @curiosityvc orq.ai/blog/orq-ai-ra…

Agent runtime, our latest product, allows teams to build enterprise-grade agents, while we handle all the infrastructure, orchestration, governance and finops across 300 models and 18 providers. orq.ai/platform/agent…

Today we're announcing a 5 million Seed round and our Agent Runtime. When we started 3 years ago, we just wanted to be on the field... Now, we are running critical AI workloads for hundreds of teams, supporting the entire AI engineering lifecycle.