Skip to content

Zhipu AI News: Latest Releases & Events

Zhipu AI (Z.ai) faces Anthropic's October allegation that it distilled US models' cyber capabilities for GLM-5.3, and is cited among open-weight providers pressuring OpenAI prices. GLM-5.3 arrived in August, surpassing Kimi K3 on many agentic coding benchmarks; June's MIT-licensed GLM-5.2 topped open weights models on the Artificial Analysis index. These followed GLM-5, a 744B-parameter model, in February.

Who is Zhipu AI? → Big AI Players

All 8 Zhipu AI entries of the AI News timeline, newest first.

2026

October 2026

2026-10-26

  • OpenAI releases GPT-6.1 Terra at half price
    OpenAI introduced GPT-6.1 Terra, a fast, efficient model for computer-based tasks like drafting proposals and editing video, priced at $1 per million input tokens and $5 per million output tokens — 50% cheaper than GPT-6.1 Sol and the fourth consecutive OpenAI price cut amid pressure from open-weight Chinese providers like Alibaba, Baidu, and Z.ai. Analysts note the escalating price war gives enterprises lower-cost options but deepens confusion over which model fits which workload, with implications for both OpenAI and Anthropic as they pursue public listings. (aibusiness.com)
    OpenAI Alibaba Baidu

2026-10-12

  • Anthropic cuts Claude distillation attacks by 90-99%
    Anthropic announced a multi-layered defense against distillation attacks on Claude—preserved thinking, a two-stage constitutional classifier, and rapid account bans—that reduced reasoning tokens extracted by distillers by roughly 90 to 99 percent. The post also alleged that Zhipu distilled US models' cyber capabilities for GLM-5.3 and that Moonshot's Kimi secretly forwarded user requests (including one linked to China's PLA) to Claude for distillation. (anthropic.com)
    Anthropic Moonshot AI

August 2026

2026-08-14

  • Z.ai releases GLM-5.3, matching frontier coding models
    Z.ai announced GLM-5.3, a roughly 750B-parameter model built on the same base as GLM-5.2 but with substantially extended post-training, which surpasses Kimi K3 on many agentic coding benchmarks and occasionally beats Claude Fable 5 and GPT-5.6-Sol despite using a third of Kimi K3's parameters. Open weights are planned on Hugging Face within two weeks under a staged rollout citing cybersecurity dual-use risks, and Nathan Lambert argues the results reflect Z.ai's post-training skill and faster release cycles rather than distillation from US labs. (interconnects.ai)

June 2026

2026-06-28

  • Open AI model ecosystem expands in diversity
    An overview of the open model ecosystem highlights a trend of increasing diversity, driven by pure model makers, Big Tech, and product companies. The review features notable recent releases including NVIDIA's Nemotron-3-Ultra, Cohere's Command A+, GLM-5.2, Zyphra's ZAYA1, and Poolside's Laguna-M.1. (interconnects.ai)
    Nvidia Cohere Zyphra

2026-06-16

  • Z.ai releases GLM-5.2 open weights LLM
    Chinese AI lab Z.ai released GLM-5.2 under an MIT license, a 753B parameter Mixture-of-Experts model with a 1 million token context window that became the leading open weights model on the Artificial Analysis Intelligence Index with a score of 51, surpassing MiniMax-M3, DeepSeek V4 Pro, and Kimi K2.6. (simonwillison.net)

March 2026

2026-03-27

February 2026

2026-02-11

  • Z.ai releases GLM-5 language model
    Z.ai announced the release of GLM-5, the latest iteration in their General Language Model series. This release signifies a major upgrade to the GLM family - with massive numbers: 744B params (40B active) based on 28,5T token in pre-training. (z.ai)

2025

December 2025

2025-12-22

  • Z.AI new GLM-4.7 AI coding model release
    Z.ai releases GLM-4.7, a coding-focused AI model showing benchmark improvements including 73.8% on SWE-bench (+5.8% over prior version) and better terminal/web UI generation capabilities.