- 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
Apertus-v1.5-70B
Swiss AI
Apache 2.0
Open language model for chat and text. 2 sizes from 8.9B to 72B parameters.
- We could not read the architecture of this model. The memory for the context is estimated.
Settings
Quality
4 bit (Q4_K_M) is the most common choice: good quality with low memory requirements.
Context length: 8k tokens
4k32k
How much text the model keeps in view at once. More context needs more memory.
Recommendation
1× H100 80 GB
The model needs about 47.4 GB of memory. Hopper, FP8 support.
- Weights
- 43.2 GB
- KV cache
- 2.7 GB
- Speed
- ~23.0 Tok./s
More suitable configurations
| Card | Quantity | Tok./s | Status |
|---|---|---|---|
| Tesla T4 16 GB | 4× | 3.2 | 2-3 weeks delivery time |
| Tesla P100 16 GB | 4× | 4.0 | 2-3 weeks delivery time |
| Tesla P40 24 GB | 4× | 3.8 | 2-3 weeks delivery time |
| Tesla V100 16 GB | 4× | 8.9 | only 1 in stock |
| Tesla V100 32 GB | 2× | 10.2 | On order, delivery time approx. 1 week |
| AMD Instinct MI50 16 GB | 4× | 6.8 | only 2 in stock |
| AMD Instinct MI50 32 GB | 2× | 7.9 | in stock |
| AMD Instinct MI60 32 GB | 2× | 7.3 | 2-3 weeks delivery time |
| AMD Instinct MI100 32 GB | 2× | 8.8 | 2-3 weeks delivery time |
| A30 24 GB | 4× | 8.6 | 2-3 weeks delivery time |
| RTX A5000 24 GB | 4× | 9.0 | 2-3 weeks delivery time |
| A40 48 GB | 2× | 9.9 | 2-3 weeks delivery time |
