# AI benchmarks
Latest news and articles about AI benchmarks
Total: 3 articles found
ByteDance Elevates 'Doubao' to 2.0 — A Productised, Multi‑variant Large Model Push with Bold Benchmark Claims
ByteDance has launched Doubao 2.0, a family of production‑oriented large models (Pro, Lite, Mini, and a Code variant) focused on efficient inference, multimodal ability and long‑chain task execution. The company claims strong benchmark performance versus rivals but faces scrutiny over IP and international review requests, highlighting the tension between rapid commercialisation and the need for independent validation and regulatory compliance.
A Night of Acceleration: Zhipu’s GLM‑5 Debuts as MiniMax and DeepSeek Race to Keep Up
Three leading Chinese AI firms unveiled near‑simultaneous upgrades that signal a shift from demo‑level coding assistants to production‑oriented, agentic systems. Zhipu launched GLM‑5 as an open‑source foundation for long‑horizon engineering tasks, while MiniMax and DeepSeek pushed product and context upgrades aimed at real‑world throughput and extended interactions.
Alibaba’s Qwen3‑Max‑Thinking: China’s Latest Push to Match Western ‘Thinking’ Models
Alibaba has launched Qwen3‑Max‑Thinking, a flagship inference model its team says matches top Western 'thinking' models on key benchmarks. The model emphasizes reasoning, instruction following and agent capabilities and is aimed at commercial integration across Alibaba's cloud and services. The announcement underscores China’s accelerating push to develop indigenous, production‑ready large language models, though benchmark claims require independent validation.