Currently listed through these providers:
Model details
GLM-5.1
GLM-5.1 represents Zhipu AI's push into next-generation agentic engineering, built as a flagship foundation model designed to handle long-horizon tasks that span hours of continuous autonomous work rather than brief interactions. The model leverages a Mixture of Experts architecture with 756 billion total parameters, activating a fraction for any given task, which allows it to scale reasoning capacity efficiently. Its design centers on autonomous planning, execution, and iterative self-improvement across extended sessions, enabling it to tackle engineering-grade projects from start to finish. The model's capabilities particularly shine in sustained coding tasks, where it outperforms its predecessor GLM-5 by a wide margin and achieves state-of-the-art results on SWE-Bench Pro.
The development lineage reflects heavy investment in post-training optimizations that distinguish it from earlier GLM versions, with the community noting that the pricing premium over GLM-5 is tied to these agentic refinements rather than architectural changes. This approach targets building autonomous agents and long-horizon coding agents that can operate independently for extended periods, making it particularly suited for complex real-world development workflows. Multiple leading AI coding tools have already adopted it as their default model, signaling confidence in its ability to handle complex engineering optimization tasks. The combination of its massive parameter scale, specialized agentic training, and extended autonomous execution window positions it as a strong choice for applications requiring sustained, multi-step reasoning without human intervention.
Quick Info
Powered by- Provider
- DInference
- Model key
- glm-5.1
- Release date
- Apr 7, 2026
- Last updated
- Apr 7, 2026
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $1.25
- Output token cost
- $3.89
Limits
- Output tokens
- 128,000 tokens
- Context window
- 200,000 tokens