Dev Meme / devmeme
Photo
Btw, gemini-2.5-pro actually worth it
The first one to seriously compete with Claude and the first one to meaningfully outperform Claude in some of aspects
Its been almost a year of total Claude's domination!
The first one to seriously compete with Claude and the first one to meaningfully outperform Claude in some of aspects
Its been almost a year of total Claude's domination!
๐คก41๐24๐ฉ7๐คฎ6๐ฅ3๐ฅฐ2๐ณ2โค1
This media is not supported in your browser
VIEW IN TELEGRAM
๐87๐ณ34๐12โค5๐คก5
I was cooking something silly and funny this night to not be so late for 4.04 celebration but wanna get a thing to a more polished and finished state (why does it sound like Iโm on daily and weโre about to miss the release date? ๐ค)
๐ค34๐คช6๐คฎ5โค3๐ฑ2๐ณ2๐คก1
Dev Meme / devmeme
Incredible start of the year
Ok, I get it, reading screeshots is hard + now nobody really cares about benchmarks:
Llama 4 Scout - 210GB / 10M context window / 17B active params x 16 experts / 109B total params
Llama 4 Maverick - 788GB / 1M context / 17B active params x 128 experts / 400B total params
Llama 4 Scout - 210GB / 10M context window / 17B active params x 16 experts / 109B total params
Llama 4 Maverick - 788GB / 1M context / 17B active params x 128 experts / 400B total params
๐คฏ29๐ฅฑ14๐คก2๐1๐ณ1
Dev Meme / devmeme
Ok, I get it, reading screeshots is hard + now nobody really cares about benchmarks: Llama 4 Scout - 210GB / 10M context window / 17B active params x 16 experts / 109B total params Llama 4 Maverick - 788GB / 1M context / 17B active params x 128 experts /โฆ
To those who are thinking "I don't have enough money to buy those H100 anyway"
It may be crazy to you at first, but stop thinking about nvidia...
It's all about Apple now!
It may be crazy to you at first, but stop thinking about nvidia...
It's all about Apple now!
๐42๐ฉ12๐คฏ6๐คก5๐ณ2๐1๐คช1
Dev Meme / devmeme
Ok, I get it, reading screeshots is hard + now nobody really cares about benchmarks: Llama 4 Scout - 210GB / 10M context window / 17B active params x 16 experts / 109B total params Llama 4 Maverick - 788GB / 1M context / 17B active params x 128 experts /โฆ
Users and companies โdomiciledโ or with a โprincipal place of businessโ in the EU are prohibited from using or distributing the models ๐
https://www.llama.com/llama4/use-policy/
https://www.llama.com/llama4/use-policy/
Please open Telegram to view this post
VIEW IN TELEGRAM
Meta Llama
Llama 4 Acceptable Use Policy
๐25๐คฉ3๐คก2โค1๐ณ1๐1