Reasoning extension of GLM-4-32B-0414
According to the model card, GLM-Z1-32B-0414 builds on GLM-4-32B-0414 with cold start, extended reinforcement learning, and further math, code, and logic training.
Open Source Model Profile · zai-org
GLM-Z1-32B-0414
GLM-Z1-32B-0414 is a 32.57B-parameter glm4 text-generation reasoning model from zai-org. According to the model card, it extends GLM-4-32B-0414 with reinforcement learning for math, code, and logic.
- Publisher
- zai-org
- Task
- text-generation
- Model type
- glm4
- License
- mit
- Library
- transformers
- Publication status
- Approved for indexing
Model overview
GLM-Z1-32B-0414 is published by zai-org as a text-generation model. The captured configuration identifies Glm4ForCausalLM, and Safetensors metadata reports 32,566,081,536 parameters. According to the model card, it is a deep-thinking reasoning model derived from GLM-4-32B-0414 with extended reinforcement learning.
Recorded capabilities
Publisher-described base-series lineage
According to the model card, the GLM-4-32B-0414 series is described as a 32B-parameter line whose base was pre-trained on 15T of data with later instruction, code, and function-calling post-training.
Documented sampling and thinking controls
According to the model card, the publisher documents temperature, top_p, top_k, and max_new_tokens values plus enforced thinking and history-trimming guidance.
Transformers usage note
According to the model card, inference code requires transformers>=4.51.3; the hub record lists transformers as the library.
Use cases in the source record
- Reasoning workflows in mathematics, code, and logic that follow the card's description of the model's deep-thinking extension.
- Conversational text-generation experiments using the card's documented sampling values and enforced-thinking format.
Limitations and unknowns
- No evaluation results were extracted from this record; comparative statements about GPT-series, DeepSeek, and GPT-4o-class models are publisher claims.
- No context-window value was extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Training, scale, and capability claims come from the publisher model card and were not independently verified.
Source and provenance
Source: zai-org/GLM-Z1-32B-0414
Captured: Unknown. Processed: 2026-09-07T19:35:33.005800+00:00.
GLM-4-Z1-32B-0414 Introduction The GLM family welcomes a new generation of open-source models, the GLM-4-32B-0414 series, featuring 32 billion parameters. Its performance is comparable to OpenAI's GPT series and DeepSeek's V3/R1 series, and it supports very user-friendly local deployment features. GLM-4-32B-Base-0414 was pre-trained on 15T of high-quality data, including a large amount of reasoning-type synthetic data, laying the foundation for subsequent reinforcement learning extensions. In the post-training stage, in addition to human preference alignment for dialogue scenarios, we also enhanced the model's performance in instruc…
F001F002F003F004F005F006F007F008F009F010F011F012F015F016