Zhipu AI Releases Open-Source GLM-4.7 Model, Challenging Top Coding AI

Zhipu AI, a Chinese artificial intelligence company, has launched and open-sourced its flagship large language model, GLM-4.7, ahead of its anticipated Hong Kong IPO. The release positions GLM-4.7 as a direct competitor to leading coding models, including Anthropic's Claude Sonnet 4.5.

Performance Benchmarks and Developer Adoption

GLM-4.7 has achieved top rankings among open-source models on several authoritative benchmarks, including the Vals Index. The Vals Index, which evaluates coding ability and complex tasks in finance and law, placed GLM-4.7 as the number one open-source model. This indicates its capability to handle high-value tasks previously dominated by closed-source models.

In developer blind tests conducted within the Design Arena, GLM-4.7 ranked second in win rate and Elo rating, trailing Google's Gemini 3 Pro Preview. It surpassed models such as Claude Opus 4.5 and GPT-5.2. The model also secured the top open-source position and sixth globally in the WebDev coding arena.

These results suggest a growing preference among developers for GLM-4.7 in practical applications, moving beyond mere parameter alignment to experiential alignment.

Developer's hands typing on a backlit keyboard, with code displayed on a monitor, symbolizing practical application.

Silicon Valley Integration and Cost Efficiency

Several mainstream AI platforms popular with overseas users, including Cline, Kilo, and Vercel, have announced integration with GLM-4.7. Fireworks, a company valued at $4 billion, also provided Day 0 support for the model.

The adoption by Silicon Valley companies is attributed to GLM-4.7's cost-effectiveness and performance. Compared to more expensive models like GPT and Claude, GLM-4.7 offers comparable coding capabilities at a lower cost, making it an attractive option for Agent tools that consume large numbers of tokens. Zhipu AI currently serves over 150,000 overseas users.

Rows of server racks in a modern data center, symbolizing the infrastructure supporting AI models.

Community Reception and Practical Applications

The release of GLM-4.7 has generated significant discussion within developer communities. Foreign users have praised its performance, with some noting that previous versions, such as GLM-4.6, were already competitive with Claude 4.5.

Examples of GLM-4.7's capabilities include generating voxel-art scenes with complex elements like pagodas and cherry blossom trees, creating skill-based Gomoku games with AI opponents, and developing browser operating systems with multiple applications and customizable features.

Zhipu AI provides detailed documentation and tutorials for integrating GLM-4.7 into popular coding environments like Claude Code. This allows developers to quickly deploy and utilize the model for various programming tasks.