Retour à Mistral Nemo
Mistral AI's 12B Nemo model co-developed with NVIDIA — strong multilingual performance with a 128K context window.
131K tokensGratuit / Poids ouvertsTransformerApache 2.0
Benchmarks
BFCL v4
Mistral-Nemo-Instruct-2407
27.6%Open-Mistral-Nemo-2407 FC — rank #78