Mistral’s new Large 4 trails some Chinese open models in independent tests
Artificial Analysis scores Mistral’s new Large 4 at 38, behind open models from Xiaomi, Z.ai, Moonshot, and DeepSeek.
Mistral has unveiled its new Large 4 model, boasting 49 billion active parameters, which is now available for a Research Public Preview. This trillion-parameter model, which pushes the frontier of open-weight performance, has been tested and ranked by independent AI benchmarking firm Artificial Analysis (AA). AA placed Mistral Large 4 as the best model outside the U.S. and China in its index, surpassing several Chinese open models in specific tests.
When compared to other open models, Mistral Large 4 trails slightly, with scores ranging from 33 to 39, depending on the model. However, it does hold an advantage in cybersecurity capabilities, scoring 50% on the Artificial Analysis Cyber Index, which is on par with GLM-5.3-Flash. This performance makes Mistral Large 4 the most intelligent model from outside the U.S. and China, according to AA's Intelligence Index.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
