mistralai/Mixtral-8x7B-Instruct-v0.1

Text Generation • Updated 19 days ago • 787k • • 3.99k

Note Oldie but goodie still my go to honestly, with systems like mine with 64GBs of RAM it's perfect balance of smartness and speed. Wish I could run 8x22B WizardLM at decent speeds and bigger than Q3

miqudev/miqu-1-70b

Updated Feb 4 • 9.64k • 973

Note Migu :O It's pretty good for what it is

alpindale/goliath-120b

Text Generation • Updated Dec 21, 2023 • 3.69k • 219

Note Big and slow, sometimes kinda smart. Maybe other models are better at this point, but I still have a soft spot for it

karakuri-ai/karakuri-lm-70b-chat-v0.1

Text Generation • Updated May 7 • 707 • 23

Note This ones only merit is being good at Japanese, everything else about it is quite poor

CohereForAI/c4ai-command-r-plus

Text Generation • Updated 6 days ago • 23.6k • 1.59k

Note This ones only merit is being good at Japanese, everything else about it is quite poor to mid but better than karakuri-lm at general stuff but worse at Japanese by a bit but still can do it

alpindale/WizardLM-2-8x22B

Text Generation • Updated Apr 16 • 116k • 363

Note Smartest model I have had the pleasure of using, somewhere in the middle of GPT-3.5 and GPT-4 but with less multi lingual abilities. Worth noting I can only judge based off of Q3 version

mistralai/Mixtral-8x22B-Instruct-v0.1

Text Generation • Updated 19 days ago • 48.4k • 629

Note I mean it's okay, WizardLM 2 kinda btfod it

meta-llama/Meta-Llama-3-70B-Instruct

Text Generation • Updated May 29 • 485k • • 1.34k

Note Decent base model

openchat/openchat-3.6-8b-20240522

Text Generation • Updated May 28 • 131k • 131

Note Maybe the best smaller model as of now

microsoft/Phi-3-mini-4k-instruct-gguf

Text Generation • Updated 20 days ago • 50.8k • 417

Note Best super small model, good for what it is but don't expect SOTA at this size