Zhipu AI's GLM-5.1 scores 58.4 on SWE-Bench Pro, surpassing GPT-5.4
Zhipu AI has released GLM-5.1, its latest flagship language model, which has achieved a groundbreaking score of 58.4 on the SWE-Bench Pro benchmark — surpassing both OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6. This marks a significant milestone for open-weight models in coding competency.
The model is specifically designed for agentic coding workflows, with the ability to work autonomously for up to 8 hours on complex engineering tasks. It demonstrated this by autonomously performing 655 iterations to optimize a vector database, achieving a 6.9x throughput improvement over the initial version.
GLM-5.1 features a 200K context window with 128K max output tokens, making it suitable for large-scale code generation and analysis. The model supports function calling, structured output, and MCP (Model Context Protocol) integration. It's available through Zhipu AI's coding plan API.
This development signals that the gap between proprietary and open-weight models is rapidly closing, particularly in coding tasks. For developers in the MENA region, GLM-5.1 offers a cost-effective alternative to Western models without compromising on quality.
GLM-5.1's performance is a clear signal that the AI coding model landscape is becoming more competitive. Open-weight models are no longer just "good enough" alternatives — they're matching or beating the best proprietary models. This has significant implications for cost-conscious developers and businesses, especially in emerging markets where access to expensive API subscriptions can be a barrier. The 8-hour autonomous work capability also raises the bar for what coding agents can achieve without human intervention.
What is SWE-Bench Pro and why does the score matter?
SWE-Bench Pro is a benchmark that tests AI models on real-world software engineering tasks — fixing bugs, implementing features, and refactoring code in actual open-source projects. A higher score means the model can handle more complex, real development work autonomously.
Is GLM-5.1 free to use?
GLM-5.1 is available through Zhipu AI's coding plan API. Some tiers offer free access with rate limits. Check the Zhipu AI platform for current pricing and availability.