Small
Model NameVersionContextSpeed*Model PageRecommendation
Meta-Llama-3-8B-Instruct1.0819250t/shttps://huggingface.co/meta-llama/Meta-Llama-3-8B-InstructGreat for general use, verbose replies, assistant focused
Awanllm-Llama-3-8B-Dolfin0.6819250t/shttps://huggingface.co/AwanLLM/Awanllm-Llama-3-8B-Dolfin-v0.6-AbliteratedGreat for general use, follows instructions better, less refusals
Awanllm-Llama-3-8B-Cumulus0.3.2819250t/shttps://huggingface.co/AwanLLM/Awanllm-Llama-3-8B-Cumulus-v0.3.2Great for RP, zero refusal, takes personalities better
Medium
Model NameVersionContextSpeed*Model PageRecommendation
Mixtral 8x7B Instruct0.13276820t/shttps://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1Great for general use, long context window
Large
Model NameVersionContextSpeed*Model PageRecommendation
Meta-Llama-3-70B-Instruct1.0819215t/shttps://huggingface.co/meta-llama/Meta-Llama-3-70B-InstructGreat for general use, SOTA open-source LLM model
Coming Soon
Model NameVersionContextSpeed*Model PageRecommendation
WizardLM-2-8x22B1.06553610t/shttps://huggingface.co/alpindale/WizardLM-2-8x22BGreat for general use, SOTA open-source LLM model, long context window
Mistral-7B-Instruct0.23276850t/shttps://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2Fast replies, long context window
*Speed is an estimation and may vary.