🚨 OpenAI recently dropped GPT-OSS: open-source reasoning models gpt-oss-120b & gpt-oss-20b. Follow along our latest deep dive blog on their architecture, benchmarks vs DeepSeek-R1, compute requirements & Hyperbolic's serverless endpoints for both models!

🚀 Alibaba's Qwen team dropped Qwen-Image, a 20B-param MMDiT model revolutionizing text-to-image gen! Native in-pixel text rendering for stunning posters, bilingual EN/CN support, excels in photorealistic/anime/styles. Dive into our technical article with full guide to deploy the model on Hyperbolic & Gradio

Research by Dr. Yu Sun: Advancing AI with Hyperbolic Compute Dr. Yu Sun, alongside researchers from Stanford, UC Berkeley, UCSD, and UT Austin, is pushing the frontier of generative AI using Hyperbolic Labs' GPU infrastructure. Two breakthrough projects: minute-long video generation and adaptive RNNs. 🧵

We're proud to be serving OpenAI's latest open-weight models: gpt-oss-20B and gpt-oss-120B, right now at Hyperbolic! 💜 Try it out here: gpt-oss-20B: gpt-oss-120B: The 120B model performance is of ~o4-mini level. The smaller model can run on most desktops and laptops while the larger model can run on high-end ones and data centers! Cheapest H100 ($1.49/hr) to run the models yourself at

Dive into @nvidia 🐰 Hopper vs ⚙️ Blackwell GPU architectures—successive leaps in compute power and interconnects for your AI workloads👇

The economics of AI are reaching a critical inflection point. According to research, training costs for frontier AI models are projected to hit more than $1 billion by 2027. @AnthropicAI's CEO Dario Amodei predicted that next year, training costs for AI models will grow to a few billion dollars per run. "In 2026, it may be above $10 billion to train a single model. By 2027, he anticipates that model companies will have ambitions to build $100 billion training clusters."

🚨 New Feature Alert from Hyperbolic 🚨 Network Storage is now live, giving you persistent storage across GPU sessions. Your data stays put even if your instance doesn’t. No more cold starts. No more re-uploading massive datasets. Let’s dive in 👇

Deploy GPU clusters fast with Hyperbolic

The White House just unveiled an AI Action Plan built around three pillars: ✍️ > Accelerating innovation > Building robust AI infrastructure > Leading in international AI diplomacy and security At Hyperbolic Labs, we're proud to support this mission by providing affordable, reliable compute that fuels AI progress. 🧵

Thank you @Alibaba_Qwen for trusting us as their partner and first third-party inference provider to serve their latest models! Check it out now at