GLM-5.3-Flash
By zai-org. 321.3 billion parameters, a context window of 1,048,576 tokens and the licence mit.
Facts
- Released
- 2026-08-25 the day the repository was first published on Hugging Face Hugging Face, read
- Licence
- mit the licence the model card declares Hugging Face, read
- Open weights
- Yes the weights are published in this Hugging Face repository Hugging Face, read
- Parameters
- 321.3 billion counted from the safetensors weight files Hugging Face, read
- Active parameters
- 25.4 billion estimate estimated from the expert counts and sizes in the configuration model configuration, read
- Knowledge cutoff
- Unknown
- Context window
- 1,048,576 tokens the maximum position embeddings in the model configuration model configuration, read
- Longest output
- Unknown
- Tool calling
- Unknown
- Structured output
- Unknown
- Reasoning controls
- Unknown
- Inputs and outputs
- text, image in, text out Hugging Face, read
- Good for
- Vision: Hugging Face files it under image-text-to-text Hugging Face, read
Prices
| Provider | Input | Output | Context | As of | Source |
|---|---|---|---|---|---|
| RunInfra | 0.100 USD | 0.400 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| sail | 0.110 USD | 0.350 USD | 1,048,576 | LiteLLM, read | |
| Baseten | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| CoreWeave | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| Deep Infra | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| Friendli | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| Hugging Face | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| Nebius Token Factory | 0.150 USD | 0.500 USD | 1,024,000 | older than 30 days; check the provider | models.dev, read |
| Novita AI | 0.150 USD | 0.500 USD | 1,048,576 | older than 30 days; check the provider | models.dev, read |
| SiliconFlow | 0.150 USD | 0.500 USD | 1,049,000 | older than 30 days; check the provider | models.dev, read |
| Together AI | 0.150 USD | 0.500 USD | 1,048,575 | older than 30 days; check the provider | models.dev, read |
| ai& | 0.150 USD | 0.500 USD | 1,048,550 | older than 30 days; check the provider | models.dev, read |
| inferenceapis | 0.200 USD | 0.660 USD | 1,048,575 | LiteLLM, read | |
| Berget.AI | 0.290 USD | 0.580 USD | 1,048,576 | models.dev, read | |
| Modal | 0.450 USD | 1.50 USD | 1,000,000 | older than 30 days; check the provider | models.dev, read |
Run it on your own hardware
Weights are the sizes of the files a source lists, or an estimate from the parameter count where none does. Memory is an estimate: weights plus KV cache plus 512 MiB and 5 percent of the weights for runtime buffers. Check it against your hardware.
| Quantization | Weights | Memory at 8,192 tokens | Memory at 32,768 tokens |
|---|---|---|---|
| UD-IQ1_S | 86.7 GiB | 92.9 GiB | 97.0 GiB |
| UD-IQ1_M | 90.9 GiB | 97.3 GiB | 101 GiB |
| UD-IQ2_XXS | 94.9 GiB | 101 GiB | 106 GiB |
| UD-Q2_K_XL | 101 GiB | 108 GiB | 112 GiB |
| UD-IQ3_XXS | 112 GiB | 120 GiB | 124 GiB |
| UD-Q3_K_XL | 137 GiB | 146 GiB | 150 GiB |
| UD-IQ4_XS | 146 GiB | 155 GiB | 159 GiB |
| UD-Q4_K_XL | 186 GiB | 197 GiB | 201 GiB |
| UD-Q5_K_XL | 224 GiB | 237 GiB | 241 GiB |
| UD-Q6_K_XL | 272 GiB | 287 GiB | 291 GiB |
| Q8_0 | 318 GiB | 335 GiB | 339 GiB |
| BF16 | 598 GiB | 629 GiB | 633 GiB |
Start it with a local runtime
- Ollama
ollama run hf.co/unsloth/GLM-5.3-Flash-GGUF:UD-IQ1_S - LM Studio
lms get unsloth/GLM-5.3-Flash-GGUF lms server start - llama.cpp server
llama-server -hf unsloth/GLM-5.3-Flash-GGUF:UD-IQ1_S --jinja - vLLM
vllm serve zai-org/GLM-5.3-Flash - SGLang
sglang serve --model-path zai-org/GLM-5.3-Flash --port 30000
Use it from your harness
Set up for ai& with the model zai-org/glm-5.3-flash. Each endpoint page has the same setup for its own address.
OpenCode
Put this in opencode.json in your project folder:
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"aiand": {
"npm": "@ai-sdk/openai-compatible",
"name": "ai&",
"options": {
"baseURL": "https://api.aiand.com/v1",
"apiKey": "{env:AIAND_API_KEY}"
},
"models": {
"zai-org/glm-5.3-flash": {
"name": "GLM-5.3-Flash"
}
}
}
}
}- OpenCode reads any OpenAI-compatible address through the @ai-sdk/openai-compatible package, and an address that speaks the Responses API through @ai-sdk/openai.
From OpenCode documentation, read .
Pi
Put this in ~/.pi/agent/models.json:
{
"providers": {
"aiand": {
"baseUrl": "https://api.aiand.com/v1",
"api": "openai-completions",
"apiKey": "$AIAND_API_KEY",
"models": [
{
"id": "zai-org/glm-5.3-flash"
}
]
}
}
}- The apiKey field can name an environment variable as $NAME.
From Pi documentation, read .
Codex
Codex speaks only the Responses API, and ai& documents no Responses address. A gateway that offers one can sit in between.
- Codex speaks the Responses API only: responses is the one supported wire API of a custom provider. Ollama and LM Studio are built in and start with --oss.
From Codex documentation, read .
Claude Code
Claude Code sends Anthropic Messages requests, and ai& documents no such address. A gateway that translates to that API can sit in between.
- Claude Code sends Anthropic Messages requests to ANTHROPIC_BASE_URL. Anthropic says it does not support routing Claude Code to models other than Claude through any gateway, so some features may not work with another model.
From Claude Code documentation, read .
Other ways to reach it
- Baseten, model
zai-org/GLM-5.3-Flash - Berget.AI, model
zai-org/GLM-5.3-Flash - DeepInfra, model
zai-org/GLM-5.3-Flash - Friendli, model
zai-org/GLM-5.3-Flash - Hugging Face, model
zai-org/GLM-5.3-Flash - Modal, model
zai-org/GLM-5.3-Flash - Nebius Token Factory, model
zai-org/GLM-5.3-Flash - Novita AI, model
zai-org/glm-5.3-flash - RunInfra, model
zai-org/GLM-5.3-Flash - SiliconFlow, model
zai-org/GLM-5.3-Flash - Together AI, model
zai-org/GLM-5.3-Flash
Published results
These are results other people published. Baltor did not run them and does not rank models by them.
- Results the maker published on the model card published by zai-org Hugging Face, read
Two more publishers are linked, never copied, because their terms do not allow it:
- Artificial Analysis publishes its own intelligence, speed and price measurements.
- OpenRouter lists current prices and uptime for each provider it routes to.
Sources of this page
Paid links
No link in this directory is a paid link or an ad, and no listing is paid for. The order and the contents of every list come from the sources named on this page.