🚀 GLM-4-32B-0414

The GLM-4-32B-0414 series models are new additions to the GLM family, boasting 32 billion parameters. Their performance rivals that of OpenAI's GPT series and DeepSeek's V3/R1 series, and they support highly user-friendly local deployment features.

✨ Features

Model Capabilities

• GLM-4-32B-Base-0414: Pre-trained on 15T of high-quality data, including a large amount of reasoning-type synthetic data, it lays the foundation for subsequent reinforcement learning extensions. After post-training with human preference alignment for dialogue scenarios and techniques like rejection sampling and reinforcement learning, it performs well in engineering code, Artifact generation, function calling, search-based Q&A, and report generation. On several benchmarks, such as code generation or specific Q&A tasks, it achieves comparable performance with larger models like GPT-4o and DeepSeek-V3-0324 (671B).
• GLM-Z1-32B-0414: A reasoning model with deep thinking capabilities, developed based on GLM-4-32B-0414 through cold start, extended reinforcement learning, and further training on tasks including mathematics, code, and logic. It significantly improves mathematical abilities and the capability to solve complex tasks compared to the base model. General reinforcement learning based on pairwise ranking feedback is also introduced during training to enhance its general capabilities.
• GLM-Z1-Rumination-32B-0414: A deep reasoning model with rumination capabilities. Unlike typical deep thinking models, it can engage in deeper and longer thinking to solve more open-ended and complex problems. It is trained through scaling end-to-end reinforcement learning with responses graded by the ground truth answers or rubrics and can use search tools during its deep thinking process to handle complex tasks, showing significant improvements in research-style writing and complex tasks.
• GLM-Z1-9B-0414: A small model (9B) trained using all the aforementioned techniques. It exhibits excellent capabilities in mathematical reasoning and general tasks, and its overall performance is top-ranked among all open-source models of the same size. Especially in resource-constrained scenarios, it achieves an excellent balance between efficiency and effectiveness, providing a powerful option for users seeking lightweight deployment.

Showcase

Animation Generation

Web Design

Model	Details
GLM-4-32B-0414	Design a drawing board that supports custom function plotting, allowing adding and deleting custom functions, and assigning colors to functions.
GLM-4-32B-0414	Design a UI for a mobile machine learning platform, which should include interfaces for training tasks, storage management, and personal statistics. The personal statistics interface should use charts to display the user's resource usage over a period. Use Tailwind CSS to style the page, and display these 3 mobile interfaces tiled on a single HTML page.

SVG Generation

Model	Details
GLM-4-32B-0414	Create a misty Jiangnan scene using SVG.
GLM-4-32B-0414	Use SVG to illustrate the training process of an LLM.

Search-Based Writing

For search-based writing tasks, we use the following system prompt to have the model respond based on search results:

Please answer the user's question based on the provided search results.

## Notes
1. Make full use of and organize the collected information, rather than simply copying and pasting, to generate a professional answer that meets the user's requirements and has depth.
2. When the provided information is sufficient, your answer should be as long as possible. Starting from the user's intention, provide a response with sufficient information and multiple perspectives.
3. Additionally, not all search results are closely related to the user's question. Please carefully甄别, filter, and utilize them.
4. The answers to objective questions are usually very brief. You can appropriately add one or two sentences of relevant information to enrich the content.
5. Please ensure that your response is beautifully formatted and highly readable. For multi-entity comparisons or listings, make good use of the list format to help the user better understand the information.
6. Unless the user requests otherwise, please keep the language of your answer consistent with the language of the user's question.
7. When appropriate, use the format such as: 【0†source】 at the end of the sentence to reference the search results.

When using, you can obtain search results through methods such as RAG or WebSearch, and wrap them in observation, for example:

[
    {
        "role": "user",
        "content": "Explore the common characteristics of children's literature, with a focus on its narrative techniques and thematic tendencies. This includes narrative techniques: common approaches in children's literature such as first-person, third-person, omniscient narrator, and interactive narration, and their influence on young readers. It also includes thematic tendencies: recurring themes in children's literature such as growth, adventure, friendship, and family, with an analysis of how these themes impact children's cognitive and emotional development. Additionally, other universal features such as the use of personification, repetitive language, symbolism and metaphor, and educational value should be considered. Please provide a detailed analytical report based on academic research, classic examples of children's literature, and expert opinions."
    },
    {
        "role": "observation",
        "content": "【{id}†{title}†{url}】\n{content}"
    },
    ...
]

For the above prompt, we use an internal or external search model to obtain the search results. Using the format shown above, we can generate the following analysis report:

📄 License

This project is licensed under the MIT license.