Z.ai: GLM 4.5V

GLM 4.5V is a vision-language model from Z.ai, released on 11 August 2025. It is built on a Mixture-of-Experts architecture with 106B total parameters and 12B activated parameters, and Z.ai describes it as a foundation model for multimodal agent applications. It sits in the text category on api14, which is where you will find it when browsing models.

The model accepts text and images as input and returns text. Its context window is 65,536 tokens, and the maximum output is 16,384 tokens. The knowledge cutoff is 31 December 2024. Because it handles both text and images, it can work over prompts that mix written instructions with visual material, and its reasoning support means it can produce intermediate reasoning alongside its answers.

Practical uses follow from that combination. You can send a screenshot, diagram, chart or photo together with a question and get a written answer, which suits document and interface understanding, visual question answering, and captioning or description tasks. The tool calling support makes it a candidate for agent style workflows where the model decides when to call a function, and the reasoning option fits problems that benefit from step by step analysis before a final answer. Long inputs are workable within the 65,536 token context, so multi image or lengthy text plus image prompts are feasible.

You call it through api14's OpenAI-compatible API using the model id "z-ai/glm-4.5v". The supported request parameters are frequency_penalty, include_reasoning, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, temperature, tool_choice, tools, top_k and top_p. If you already use an OpenAI style client, you point it at the api14 base URL, set the model field to that id, and pass any of these parameters as needed. Use max_tokens to cap output length, temperature, top_p and top_k to steer sampling, and the tools and tool_choice pair when you want function calling. The reasoning and include_reasoning parameters control whether reasoning content is requested and returned.

Keep the limits in mind when choosing. Output is capped at 16,384 tokens, and the model returns text only, so it cannot generate images. If your task is purely text and does not benefit from image input or reasoning, a text only model may be a simpler fit. If you need image generation rather than image understanding, this model is not the right choice.