- 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
Devstral-2-123B-Instruct-2512
Mistral AI
Apache 2.0
Open model for programmers. 2 sizes from 24B to 125B parameters.
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× Tesla V100 32 GB
The model needs about 79.5 GB of memory. Passively cooled, fan and 8-pin EPS cable required.
- Weights
- 75.0 GB
- KV cache
- 3.0 GB
- Speed
- ~5.2 Tok./s
More suitable configurations
| Card | Quantity | Tok./s | Status |
|---|---|---|---|
| Tesla T4 16 GB | 8× | 1.5 | 2-3 weeks delivery time |
| Tesla P100 16 GB | 8× | 1.9 | 2-3 weeks delivery time |
| Tesla P40 24 GB | 4× | 2.2 | 2-3 weeks delivery time |
| Tesla V100 16 GB | 8× | 4.1 | only 1 in stock |
| Tesla V100 32 GB | 4× | 5.2 | On order, delivery time approx. 1 week |
| AMD Instinct MI50 16 GB | 8× | 3.2 | only 2 in stock |
| AMD Instinct MI50 32 GB | 4× | 4.0 | only 2 in stock |
| AMD Instinct MI60 32 GB | 4× | 3.7 | 2-3 weeks delivery time |
| AMD Instinct MI100 32 GB | 4× | 4.5 | 2-3 weeks delivery time |
| A30 24 GB | 4× | 5.0 | 2-3 weeks delivery time |
| RTX A5000 24 GB | 4× | 5.2 | 2-3 weeks delivery time |
| A40 48 GB | 2× | 5.7 | 2-3 weeks delivery time |
