Qwen/Qwen2.5-72B-Instruct

Text Generation • Updated Sep 25 • 228k • • 630

Note Current best for Japanese translation

google/gemma-2-27b-it

Text Generation • Updated Aug 27 • 182k • 473

Note Really nice mix of speed and intelligence, somewhat good in Japanese but mixed results maybe

princeton-nlp/gemma-2-9b-it-SimPO

Text Generation • Updated Aug 2 • 16.3k • 137

Note Current best fast model, sometimes will become weird and spit out random stuff not in English but overall good. Great world knowledge for size, somewhat poor coding(?) and low context but still good for RAG

allenai/Llama-3.1-Tulu-3-70B

Text Generation • Updated 30 days ago • 3k • 44

Note Best overall model