Teams running long, unattended coding jobs can try it under MIT; Z.ai's eight-hour claim is worth testing yourself before relying on it.
Kenning's take
What it does
GLM-5.1 is a Z.ai text model designed for long tasks, which Z.ai says can work on its own for up to eight hours in one run. It targets planning, bug fixing and repeated refinement in software work. It is available through the Z.ai API, the GLM Coding Plan, OpenRouter, and as MIT-licensed open weights.
- Long autonomous runs of up to 8 hours
- Planning, execution and iterative refinement
- Improved tool use over long tasks
- Available on OpenRouter
- MIT open-source licence
Who it suits. Developers who run long, unattended coding agent jobs.
Key facts
- Category
- AI models
- Made by
- Z.ai (Zhipu AI)
- Based in
- China
- Released
- 7 April 2026
- How to get it
- Z.ai API (model glm-5.1), GLM Coding Plan, OpenRouter, or self-hosted from open weights
- Works on
- API, open weights (Hugging Face)
- In the UK
- Yes
- API price
- $1.40 per million input tokens, $4.40 per million output tokens
Pricing
| Plan | Price | Billed |
|---|---|---|
| GLM Coding Plan | $from 18 | month |
FAQ
When was GLM-5.1 released?
Z.ai listed GLM-5.1 in its release notes on 7 April 2026.
How long can GLM-5.1 work on a task?
Z.ai says up to eight hours in a single run.
Facts from the maker's own pages and the sources listed. Checked 2026-10-11.
Similar tools
ACE-Step 1.5
Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.
AlphaGenome Atlas
Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.
Amazon Nova 2
Amazon's second family of Nova AI models, covering reasoning, speech conversation and combined text-and-image generation.
Cisco Antares
Cisco's small open-weight models that find where known security vulnerabilities sit in a codebase, running locally.