Intensifying Competition Among Domestic AI Large Models: Zhipu GLM-5 and MiniMax 2.5 Launch Together

The domestic large model field has seen a surge during the Spring Festival. After DeepSeek released a new model, Zhipu AI's GLM-5 and MiniMax2.5 were also launched today.

As Zhipu AI's latest flagship product, GLM-5 is now available on the official website z.ai. The model is positioned as a versatile model for dialogue, programming, and intelligent agents, adopting a technical approach similar to DeepSeek, including **Sparse Attention (DSA) and Multi-Token Prediction (MTP)** technologies. GLM-5 has a total of 745B parameters, about twice that of its predecessor GLM-4.7. Its architecture includes 78 hidden layers and 256 experts (8 are activated each time, with about 44B activated parameters), supporting a maximum context window of 202K Token.

At the same time, MiniMax2.5 has also made its debut on its official website. Notably, this is just over a month after the release of its 2.2 version, showing an extremely fast iteration speed. Both models have focused on enhancing programming capabilities and intelligent agent performance, aiming to compete with the world's top AI levels.

Domestic Large Models Start a Price Hike! Zhipu's GLM-5 Fires the First Shot: Overseas API Prices Rise by Up to Double, Has the Era of Low Prices Come to an End?

Zhipu's GLM-5 model goes global and raises prices, marking the first significant price increase for domestic large models, signaling a shift in the industry's commercialization logic. The price adjustments differ between domestic and international markets, with overseas increases far exceeding those domestically.

Rivaling Claude 4.5! Silicon-based Flow Launches High-speed Version GLM-5, Domestic Large Model Secures Fourth Place Globally

The domestic large model GLM-5 achieved significant breakthroughs in early 2026, ranking fourth in the authoritative list Artificial Analysis globally after being open-sourced, with a score comparable to Claude Opus4.5. Its core innovations include: expanding parameter scale to 744B and pre-training data reaching 28.5T; integrating DeepSeek sparse attention mechanism, which maintains long-text understanding capabilities while reducing deployment costs; and excelling in programming and engineering fields.

Zhipu Launches GLM-5: Entering the Agentic Ready Era from Code Generation to Engineering Construction

GLM-5, a new open-source base model, achieves a paradigm shift from code generation to independently handling complex system engineering. It ranks fourth globally and first among open-source models on authoritative benchmarks, matching top-tier performance with exceptional engineering capabilities and a user experience close to Claude Opus4.....

GLM-5 Leaked by Accident? Reusing DeepSeek Architecture, Performance Shocks, Market Value Soars 200%, Securing Top Position in Domestic AI

The domestic large model sector saw a breakthrough during the 2026 Spring Festival. Zhipu AI released its new large model GLM-5, triggering a strong reaction in the capital market, with a stock price surge of 200% and a market value reaching 150 billion HKD. Previously, the model was anonymously introduced as "Pony Alpha", showing outstanding coding capabilities.

Intensifying Competition Among Domestic AI Large Models: Zhipu GLM-5 and MiniMax 2.5 Launch Together

Related Recommendations

Domestic Large Models Start a Price Hike! Zhipu's GLM-5 Fires the First Shot: Overseas API Prices Rise by Up to Double, Has the Era of Low Prices Come to an End?

Rivaling Claude 4.5! Silicon-based Flow Launches High-speed Version GLM-5, Domestic Large Model Secures Fourth Place Globally

Another Breakthrough in Domestic Computing Infrastructure! Moortu MTT S5000 Completes Full-Process Compatibility with Zhipu GLM-5 Large Model

Zhipu Launches GLM-5: Entering the Agentic Ready Era from Code Generation to Engineering Construction

GLM-5 Leaked by Accident? Reusing DeepSeek Architecture, Performance Shocks, Market Value Soars 200%, Securing Top Position in Domestic AI