Mistral Large 3 (675B) vs Mistral Small 4 (119B-A6B)
Mistral Large 3 (675B) and Mistral Small 4 (119B-A6B) are compared here on the fields this catalogue records rather than on a benchmark score. Context window: 262,144 on both cards. Parameters: 675B for Mistral Large 3 (675B), 119.4B for Mistral Small 4 (119B-A6B). License: apache-2.0 on both cards. Modalities: text, vision on both cards.
Pick Mistral Large 3 (675B) if the larger recorded parameter count (675B against 119.4B) fits the hardware you have.
Pick Mistral Small 4 (119B-A6B) if it is the model you already intended to deploy, and nothing recorded above rules it out.
Pick neither if you need an endpoint at all: neither model has one in this catalogue.
Where the records differ
- Parameters
- Mistral Large 3 (675B): 675B · Mistral Small 4 (119B-A6B): 119.4B
- Languages
- Mistral Large 3 (675B): en, fr, es, de, it, pt · Mistral Small 4 (119B-A6B): en, fr, de, es, it, pt
| Mistral Large 3 (675B) | Mistral Small 4 (119B-A6B) | |
|---|---|---|
| Hosted by LLM EU | Catalogue only | Catalogue only |
| Region | — | — |
| Price per million tokens | — | — |
| Context window | 262k | 262k |
| License | apache-2.0 | apache-2.0 |
| Languages | en, fr, es, de, it, pt | en, fr, de, es, it, pt |
| Tasks | reasoning, coding, multilingual-support, vision, summarization | coding, reasoning, vision, multilingual-support, summarization |
| Verification source | — | — |
Where two models differ, this table shows the difference and stops. LLM EU publishes no winner badge without a documented score for that task, and a missing figure stays missing instead of becoming a zero.