
A Stellar Breakthrough: 🌌 Chinese startup Z.ai (formerly Zhipu) just launched GLM-4.5, an open-source AI model family that’s shaking the $64B AI market, per Reuters. Undercutting DeepSeek’s pricing, GLM-4.5 blends reasoning, coding, and agentic tasks, rivaling OpenAI’s o3 and xAI’s Grok 4. Released at the World AI Conference 2025, this galactic contender is redefining open-source AI. Let’s dive into this starry launch!
What’s in the Stars?
GLM-4.5, with 355B parameters (32B active), uses a Mixture-of-Experts architecture and a hybrid thinking/non-thinking mode to balance speed and depth, per VentureBeat. It ranks third globally across 12 benchmarks, scoring 90.6% in agentic tasks, 98.2% on MATH 500, and 64.2% on SWE-bench Verified, outperforming Claude 4 Opus in web tasks. The lighter GLM-4.5-Air (106B parameters, 12B active) ensures efficiency on just eight Nvidia H20 chips.
Priced at $0.11 per million input tokens and $0.28 per output, it’s 87% cheaper than DeepSeek’s R1. Z.ai’s open-source ‘slime’ training framework, available on Hugging Face, empowers developers to build custom models. X posts, like @TechBit, praise its cost-efficiency but flag privacy concerns due to Z.ai’s U.S. Entity List status.
Future Starfield
GLM-4.5’s affordability and performance could democratize AI for startups and developing nations, challenging Western giants by 2030. As China’s 1,509 LLMs lead globally, per Xinhua, Z.ai’s open-source push pressures OpenAI and Anthropic to innovate. But will data privacy and geopolitical tensions dim its shine? UrviumAI tracks this cosmic race.
🌠 Ride the Cosmic Wave! Subscribe at UrviumAI.com for AI updates.
Read more about “China’s AI Action Plan: A Galactic Call for Global Unity“
Jigar Chaudhary is the Editor-in-Chief at UrviumAI, where he oversees coverage of artificial intelligence news, tools, and in-depth studies. With over 5 years of experience analyzing AI and robotics, he focuses on maintaining high editorial standards, accurate reporting, and clear explanations to help readers understand how AI is shaping the future.




Pingback: Meta’s $1B Talent Grab at Thinking Machines Lab: AI Clash