melvincarvalho on Nostr: Open source continues to make impressive strides across the board! Meta has announced ...
Open source continues to make impressive strides across the board! Meta has announced Llama 3.3 (70B), reportedly outperforming GPT-4o while being 25x cheaper. Quite remarkable: 70B parameters, better than GPT-4o, and said to rival the 405B version of Llama 3.1.
Published at
2024-12-10 15:12:46Event JSON
{
"id": "040f1ac9d01a3f50606c0df7752a636ef861760c6625dd2e1a745730a4d80f12",
"pubkey": "de7ecd1e2976a6adb2ffa5f4db81a7d812c8bb6698aa00dcf1e76adb55efd645",
"created_at": 1733843566,
"kind": 1,
"tags": [
[
"imeta",
"url https://image.nostr.build/128c27c7448a0b098a7f1d57d972a70b634056d25cc8e406bc2f4fa363e63f43.png",
"m image/png",
"x daf04a9f7e2f8e76bb0535ad8fcdb72f399e089f329b2edea4ab9b09dce3f5c9",
"ox 128c27c7448a0b098a7f1d57d972a70b634056d25cc8e406bc2f4fa363e63f43",
"size 541437",
"dim 1491x1389",
"blurhash LSRWC?%#Vq~Ux]ofj?oeRkbHj@s:"
]
],
"content": "Open source continues to make impressive strides across the board! Meta has announced Llama 3.3 (70B), reportedly outperforming GPT-4o while being 25x cheaper. Quite remarkable: 70B parameters, better than GPT-4o, and said to rival the 405B version of Llama 3.1. \n\nhttps://image.nostr.build/128c27c7448a0b098a7f1d57d972a70b634056d25cc8e406bc2f4fa363e63f43.png",
"sig": "d1984e2caeb2209315cfb2e286caf66080a4894616b3957ea9ac6d7a293a349c0c6f1edfebbdff11e98680284bcc89a29ed8536743b9c50765b188cf022bdd88"
}
