MODELS
GLM-4.5
Open-weight Chinese frontier model with vision and tools at budget pricing.
Specs
- Context window
- 128,000
- Max output
- 8,192
- Modalities
- text, image
- Tool use
- ✓
- Vision
- ✓
- Streaming
- ✓
- License
- apache-2.0
- Released
- 2025-07-01
Pricing
- Input / 1M
- $0.50
- Output / 1M
- $1.50
Cost estimate
GLM-4.5 is Zhipu's flagship open-weight model (Apache 2.0), handling text and images with native tool use across a 128K context. It targets builders who want a capable Chinese-strong model they can self-host or call cheaply at $0.5/$1.5 per million tokens. Strong on Chinese reasoning, agentic workflows, and code; output is capped at 8K tokens.
Editor's verdict
Pick GLM-4.5 when you need a self-hostable model with solid Chinese performance and tool use, or when API budget matters more than squeezing out the last few points on English benchmarks. It's competitive with Qwen and DeepSeek in the open-weight tier but trails GPT-5 and Claude on hard reasoning. The 8K output cap can bite on long code generation or report tasks.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29