Mistral Large 2
Mistral Large 2 is released and it outperforms Llama-3.1 405B and Opus-3 on many benchmarks
Blog: https://lnkd.in/e8HKWHnX
Weights: https://lnkd.in/eT-tzMD5
Code & Reasoning:
Trained with 80+ programming languages
Mistral Large 2 (123B) comparable to GPT-4o, Opus-3 and Llama-3 405B at coding benchmarks
As compared to Mistral Large 1, significantly reduced hallucinations, improved reliability
Instruction Following:
Mistral Large 2 is particularly better at following precise instructions and handling long multi-turn conversations.
On Wild Bench, Arena Hard & MT Bench:
Outperforms Llama 3.1 405B & Opus-3
Comparable to Sonnet-3.5 and GPT-4o
Multilinguality:
Mistral Large 2 is trained on a large proportion of multilingual data.
Excels in English, French, German, Spanish, Italian, Portuguese, Chinese, Japanese, Korean, Arabic, and Hindi.
Outperforms Llama-3.1 70B, Comparable to Llama-3.1 405B
Access and Customization:
Available today on La Plateforme (https://lnkd.in/ecugAuKW) and Le Chat (https://lnkd.in/eAG-z4P8)
Available on Google Vertex AI, Azure AI Studio, Amazon Bedrock & IBM Watsonx
Also available for fine-tuning on La Plateforme along with Mistral Nemo 12B and Codestral 22B
#mistral #mistralai #llms #generativeai #llama3