- 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
Mistral-Large-3-675B-Instruct-2512
Mistral AI
GermanApache 2.0
Open language model for chat and text. 3 sizes from 119.4B to 675B parameters, partly mixture-of-experts (only a part of the parameters is active per token).
- 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
8× H100 80 GB
The model needs about 424.6 GB of memory. Hopper, FP8 support.
- Weights
- 405.0 GB
- KV cache
- 18.1 GB
- Speed
- ~1.6 Tok./s
More suitable configurations
| Card | Quantity | Tok./s | Status |
|---|---|---|---|
| A100 80 GB | 8× | 1.1 | 2-3 weeks delivery time |
| H100 80 GB | 8× | 1.6 | only 1 in stock |
