- Chat
- Vision
- Video
GLM 5.3 Flash
Long-context chat, coding and tool workflows with GLM 5.3 Flash.
GLM model family
About this model
GLM 5.3 Flash is available through the Arnict API with a 1,048,576-token context window and up to 131,072 output tokens. Input, cached input and output are priced separately. Check the live service status before sending requests.
- Parameters
- 320B18B active per token
- Native context
- 1,048,5761,048,576 tokens served by Arnict.
- Maximum output
- 131,072Hard limit
Capabilities
- Chat completionsUse this model with the chat-completions API when it is marked Available.
- Native input modalitiesText, Image, Video input are listed in the published catalog metadata.
- Long-context input1,048,576 tokens · 1,048,576 tokens served by Arnict.
- Defined output limit131,072 tokens
Published limits
Published model-card facts, kept separate from live availability.
| Metric | Published value | Note |
|---|---|---|
| Arnict API output limit | 131,072 tokens | Maximum output accepted by the Arnict API for this model |
| Parameters | 320B total / 18B active | Upstream model card |
| Context window | 1,048,576 tokens | 1,048,576 tokens served by Arnict. |
| Maximum output | 131,072 tokens hard limit | Upstream model card |
| Weights / activations | NVFP4 | Published format |
- Native multimodal model for text, image, and video
- 288 routed experts; 8 selected per token; 1 shared expert
- 45 layers
- Manifold-Constrained Hyper-Connections (mHC)
- Thinking modes: off, low, high, max
Use zai/glm-5.3-flash-uncensored with an OpenAI-compatible client now.
Call it with the OpenAI SDK
Use the complete model ID with the OpenAI SDK. Set your Arnict API key as an environment variable before running the example.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.arnict.com/v1",
apiKey: process.env.ARNICT_API_KEY,
});
const completion = await client.chat.completions.create({
model: "zai/glm-5.3-flash-uncensored",
messages: [
{ role: "user", content: "Explain continuous batching in one paragraph." },
],
});
console.log(completion.choices[0].message.content);Related models
GLM
GLM 5.3
A longer-term GLM family target for the Arnict catalog.
DeepSeek
DeepSeek V4.1 Flash
Multimodal mixture-of-experts model: 552B backbone parameters, with 8B active during prefill and 16B during decoding.
Qwen
Qwen 3.6 35B A3B
Vision-language mixture-of-experts model: 35B total, 3B active parameters.