- 3 % discount on prepayment
- Insured shipping from Switzerland
- Sustainable: refurbished instead of new
- Warranty on all cards
- 3 % discount on prepayment
- Insured shipping from Switzerland
- Sustainable: refurbished instead of new
- Warranty on all cards
← Back to overview
GLM-5.3-Flash
Z.ai (Zhipu)
MIT
Open language model for chat and text. 3 sizes from 31.2B to 753.3B parameters, partly mixture-of-experts (only a part of the parameters is active per token).
- Mixture-of-experts: 321.3B parameters in total, of which 25.4B active per token. Memory requirement depends on all parameters, speed on the active ones.
Settings
Quality
4 bit (Q4_K_M) is the most common choice: good quality with low memory requirements.
Context length: 8k tokens
4k256k
How much text the model keeps in view at once. More context needs more memory.
Recommendation
4× H100 80 GB
The model needs about 194.4 GB of memory. Hopper, FP8 support.
- Weights
- 192.8 GB
- KV cache
- 0.1 GB
- Speed
- ~54.0 Tok./s
More suitable configurations
| Card | Quantity | Tok./s | Status |
|---|---|---|---|
| Tesla V100 32 GB | 8× | 20.7 | only 6 in stock |
| AMD Instinct MI50 32 GB | 8× | 15.9 | only 2 in stock |
| AMD Instinct MI60 32 GB | 8× | 14.6 | 2-3 weeks delivery time |
| AMD Instinct MI100 32 GB | 8× | 17.7 | 2-3 weeks delivery time |
| A40 48 GB | 8× | 19.9 | 2-3 weeks delivery time |
| RTX A6000 48 GB | 8× | 22.2 | 2-3 weeks delivery time |
| Quadro RTX 8000 48 GB | 8× | 16.4 | 2-3 weeks delivery time |
| L40 48 GB | 8× | 24.4 | 2-3 weeks delivery time |
| A100 40 GB | 8× | 24.6 | 2-3 weeks delivery time |
| A100 80 GB | 4× | 38.5 | 2-3 weeks delivery time |
| H100 80 GB | 4× | 54.0 | only 1 in stock |
