3 ms·
^ This, as well as there was a lot of confusion over DeepSeek when it was released, the reasoning models were built on other models, inter alia Qwen (Chinese) a
by refulgentis 2y ago
^ This, as well as there was a lot of confusion over DeepSeek when it was released, the reasoning models were built on other models, inter alia Qwen (Chinese) and Llama (US). So one's mileage varied significantly