Skip to content

ModelsZ.ai

Z.ai releases GLM-4.5 and GLM-4.5-Air, open hybrid reasoning models for agents and coding

GLM-4.5 at a glance

GLM-4.5 (355B total, 32B active) and GLM-4.5-Air (106B total, 12B active) offer thinking and instant modes, 128K context and native function calling, with open weights.

What changed with GLM-4.5

On July 28, 2025, Z.ai introduced GLM-4.5 and GLM-4.5-Air. GLM-4.5 has 355 billion total and 32 billion active parameters; Air has 106 billion total and 12 billion active. Both are hybrid reasoning models with a thinking mode for complex reasoning and tool use and a non-thinking mode for instant replies.

Z.ai said GLM-4.5 ranked third overall across 12 agentic, reasoning and coding benchmarks against models from OpenAI, Anthropic, Google, xAI and others, with a 128K context window and native function calling. In its own tests inside Claude Code, GLM-4.5 reported a 90.6% tool-calling success rate. The models are on the Z.ai API with open weights on Hugging Face and ModelScope; the Hugging Face repository lists an MIT license.

GLM-4.5 in numbers

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Context windowas of · Source for Context window: z.ai (opens in a new tab)
128K tokens

How it compares

Other models in the same category that we have covered, using each model's latest recorded value. (Models)

Context window (tokens) · GLM-4.5higher is better

Where to compare GLM-4.5

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why GLM-4.5 matters for businesses

For businesses, GLM-4.5 and especially the smaller Air model added capable, permissively licensed options for agent workflows that can run on more modest hardware than trillion-parameter models.

Sources for this story

  1. Official
  2. zai-org/GLM-4.5 (opens in a new tab)

    Hugging Face · huggingface.co

    Official
  3. GLM-4.5 model page (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #llm
  • #zai
  • #glm
  • #open-weights
  • #agents

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.