The smallest local model that can match/beat gpt-4o-mini

Posted by ihatebeinganonymous@reddit | LocalLLaMA | View on Reddit | 16 comments

Hi. What do you think is the smallest Open model, as of Spring 2026, that can match/beat OpenAI GPT 4o Mini in the broadest sense possible?

And in a more general sense, what is the minimum number of parameters for a 2026 model to be able to match a 2024 SotA model such as 4o-mini or 4-turbo?

In case quantisation matters, assume 4 bits per weight.

Thanks

[-]

gaspoweredcat@reddit

hmm apparently gemma4 comes close, seems almost head to head with the 31b according to this

https://www.gemma4.wiki/comparison/gemma-4-vs-gpt-4o-mini

i cant say if its right or not but i dont have reason to disbelieve it really

[-]

RedParaglider@reddit

FYI your site redirects to what looks to be a porn site.

[-]

gaspoweredcat@reddit

what the hell? it took me to P&O ferries when i clicked, it was just a blog page when i added it

[-]

ttkciar@reddit

Violates Rule One: Search before asking. There is a thread specifically for this: https://old.reddit.com/r/LocalLLaMA/comments/1sknx6n/best_local_llms_apr_2026/

[-]

ihatebeinganonymous@reddit (OP)

It's really not the same. But you are the boss :D

[-]

ttkciar@reddit

I get that it's not a perfect fit, but moderators are trying to encourage use of that thread to find the model you want. Otherwise the sub gets inundated with "which model for me?" posts.

[-]

VoiceApprehensive893@reddit

gemma 4 26b beats the full 4o

[-]

ihatebeinganonymous@reddit (OP)

Even knowledge-wise?

[-]

ihatebeinganonymous@reddit (OP)

Even knowledge-wise?

[-]