flash on Nostr: β‘οΈπ«π·π€ NEW - Mistral unveils Large 4, nicknamed βLe Chonk,β with ...
β‘οΈπ«π·π€ NEW - Mistral unveils Large 4, nicknamed βLe Chonk,β with 1,000 billion parameters, 49 billion of which are active.
The multimodal model claims to deliver the best performance among American and European open-weight models on aggregated benchmarks.
In particular, it is said to be at the forefront in cybersecurity, industry, and finance, and to outperform leading closed models in visual understanding.
Developed entirely in Europe, it is available today via API on Mistral Cloud. The open weights are scheduled to be released in late October.
{
"id":"50d2c0bb74f2f39a71e048d4dde42fa759995063a9eabd59457b3b4b054b8e03",
"pubkey":"4d7842051782e0d3feb034d150adc2b6bae4ee3b49786793bffa468b6f5b96b3",
"created_at":1791355841,
"kind":1,
"tags": [
[
"client",
"Primal iOS"
]
],
"content":"β‘οΈπ«π·π€ NEW - Mistral unveils Large 4, nicknamed βLe Chonk,β with 1,000 billion parameters, 49 billion of which are active.\n\nThe multimodal model claims to deliver the best performance among American and European open-weight models on aggregated benchmarks.\n\nIn particular, it is said to be at the forefront in cybersecurity, industry, and finance, and to outperform leading closed models in visual understanding.\n\nDeveloped entirely in Europe, it is available today via API on Mistral Cloud. The open weights are scheduled to be released in late October. \nhttps://24242.io/a78ccf92051e5d9e57bc6e70a8061cd43828302f834cc498915b8a38ee42ebb8.mp4",
"sig":"feb4d4b7a6b2446f27d35a33cb82dfe6dedd0ce100758e53cbbfb08097becc1224f23302d3b8ffd68eda4cdf6124b8ffcbf4dc44c5cf419c30dff2bfbca16d79"
}