Will top open-weight LLMs in 2025 reason opaquely?

Will Llama 4 be the best LLM in the chatbot arena?

Will a flagship (>60T training bytes) open-weights LLM from Meta which doesn't use a tokenizer be released in 2025?

Was Llama 4 sabotaged?

Will Llama 4 Behemoth top the lmarena?

How many active parameters will the largest Llama 3 have?

Will Llama-3 (or next open Meta model) be obviously good in its first-order effects on the world?

Will SB 1047 become law? Will Meta open source Llama 4 (or equivalent)