GLM-5.3-Flash
GLM-5.3-Flash is a new mixture-of-experts base model from Z.ai with about 320B total and 18B active parameters, released in August 2026 under MIT. It is the first natively multimodal GLM-5 model and combines sparse and linear attention to cut long-context serving cost. The API accepts text, images, video and files with a 1M-token context.
Specifications
- Context window
- 1Mabout 786,000 words, or 1,570 pages
- Max output
- 128Ktokens per response
- Input price
- $0.15per 1M tokens
- Output price
- $0.50per 1M tokens
- Cached input
- $0.03per 1M tokens
- Weights
- OpenMIT
- Parameters
- 321B total, 18B activeMoE
- Input types
- Text, Images, Video
- Output types
- Text
- Reasoning mode
- Yes
- Tool calling
- Yes
- Released
- Aug 26, 2026Current
Z.ai API price for GLM-5.3-Flash; a faster GLM-5.3-FlashX tier costs $0.37 / $0.075 cached / $1.25.
What it costs in practice
| Workload | Monthly cost at list price |
|---|---|
| Support chatbot30,000 conversations a month, 2,000 input and 400 output tokens each | $15 |
| Document summaries2,000 documents a month, 12,000 input and 800 output tokens each | $4.40 |
| Coding agent1,000 tasks a month, 60,000 input and 6,000 output tokens each | $12 |
Standard rates, no caching or batch discounts. Run your own numbers.
Strengths
- MIT license
- Native image, video and file input
- Hybrid sparse/linear attention for cheaper long contexts
Trade-offs to weigh
- Thinking cannot be turned off
- 320B total parameters still require a multi-GPU server
Where it fits
Z.ai (Zhipu AI) positions GLM-5.3-Flash for:
Where you can use it
- Hugging Face
- Ollama
- API
- Together AI
- Fireworks
Run it yourself
Released under MIT. Approximate memory needed just to load the weights:
| Precision | Memory for weights |
|---|---|
| 16-bit (BF16/FP16) | 642 GB |
| 8-bit (FP8/INT8) | 321 GB |
| 4-bit (Q4_K_M, about 4.85 bits) | 195 GB |
This is a mixture-of-experts model: all 321B total parameters must sit in memory, but only about 18B are used per token, so it runs faster than a dense model of the same size.
Add memory for the KV cache (it grows with context length and concurrent users) and 1-2 GB of runtime overhead. Size it precisely in the VRAM calculator.
Alternatives to consider
| Model | Tier | Context | Input / output per 1M | |
|---|---|---|---|---|
| Gemma 4 31BGoogle | Efficient | 256K | $0.20 / $0.40 | Compare |
| Mistral Small 4Mistral AI | Efficient | 256K | $0.15 / $0.60 | Compare |
| Ministral 3 14BMistral AI | Efficient | 256K | $0.20 / $0.20 | Compare |
| Granite 4.2 30BIBM | Efficient | 128K | $0.16 / $0.65 | Compare |
Frequently asked questions
How much does GLM-5.3-Flash cost?
GLM-5.3-Flash lists at $0.15 per million input tokens and $0.50 per million output tokens, with cached input at $0.03 (checked October 10, 2026). Z.ai API price for GLM-5.3-Flash; a faster GLM-5.3-FlashX tier costs $0.37 / $0.075 cached / $1.25. Confirm current rates on the provider's pricing page before you commit.
What is the context window of GLM-5.3-Flash?
GLM-5.3-Flash accepts up to 1,048,576 tokens of context (about 786,000 words, or 1,570 pages), and it can write up to 128K tokens in a single response. Long prompts cost more and can be slower, so most apps send only what the task needs.
Is GLM-5.3-Flash open source?
GLM-5.3-Flash is an open-weight model released under the MIT license. You can download the weights from Hugging Face (zai-org/GLM-5.3-Flash) and run or fine-tune them yourself, within the license terms. Open weights isn't the same as fully open source: training data and code may not be published.
Can I run GLM-5.3-Flash locally?
Yes, if you have enough memory. The weights alone need about 195 GB at 4-bit quantization and about 642 GB at 16-bit, plus room for the KV cache. Use the VRAM calculator on LLModel.com to check a specific GPU or Mac.
What is GLM-5.3-Flash best for?
Z.ai (Zhipu AI) positions GLM-5.3-Flash for agents, multimodal, high volume, long documents. It accepts text, images, video as input, has a reasoning (thinking) mode and supports tool calling. Test it on a sample of your own tasks before you choose it for production.
Sources
- huggingface.co/zai-org/GLM-5.3-Flash
- docs.z.ai/guides/overview/pricing
- docs.z.ai/release-notes/new-released
- docs.z.ai/guides/llm/glm-5.3-flash
- huggingface.co/zai-org/GLM-5.3-Flash/raw/main/config.json
- huggingface.co/api/models/zai-org/GLM-5.3-Flash
- docs.together.ai/docs/serverless-models
- docs.fireworks.ai/serverless/pricing
- ollama.com/library
Facts checked against these pages on October 10, 2026. Spot an error? Tell us.
Still choosing a model?
Send us your use case and constraints. We'll come back with a shortlist and an evaluation plan you can run on your own data.
- Free and vendor-neutral: no model provider pays for placement.
- Your details go only to partners you agree to.